path
stringlengths 13
17
| screenshot_names
sequencelengths 1
873
| code
stringlengths 0
40.4k
| cell_type
stringclasses 1
value |
|---|---|---|---|
16127029/cell_29 | [
"text_plain_output_1.png"
] | import matplotlib.pyplot as plt
import pandas as pd # data processing, CSV file I/O (e.g. pd.read_csv)
data = pd.read_csv('../input/loan.csv', low_memory=False)
data = data[(data.loan_status == 'Fully Paid') | (data.loan_status == 'Default')]
data['target'] = data.loan_status == 'Fully Paid'
data.shape
x = data.loan_amnt
term_groups = data.groupby('term')
term_groups['int_rate'].mean()
grade_groups = data.groupby('grade')
grade_groups['int_rate'].mean()
X = pd.get_dummies(data[['term', 'verification_status', 'purpose', 'policy_code', 'loan_amnt', 'funded_amnt', 'funded_amnt_inv', 'int_rate', 'emp_length', 'addr_state']])
count_class_0, count_class_1 = data.target.value_counts()
data_class_0 = data[data['target'] == 1]
data_class_1 = data[data['target'] == 0]
data_class_0_under = data_class_0.sample(count_class_1)
data_test_under = pd.concat([data_class_0_under, data_class_1], axis=0)
print('Random under-sampling:')
print(data_test_under.target.value_counts())
data_test_under.target.value_counts().plot(kind='bar', title='Count (targe t)') | code |
16127029/cell_26 | [
"text_plain_output_1.png"
] | from sklearn.ensemble import RandomForestClassifier
from sklearn.metrics import accuracy_score
from sklearn.model_selection import train_test_split
import matplotlib.pyplot as plt
import numpy as np # linear algebra
import pandas as pd # data processing, CSV file I/O (e.g. pd.read_csv)
data = pd.read_csv('../input/loan.csv', low_memory=False)
data = data[(data.loan_status == 'Fully Paid') | (data.loan_status == 'Default')]
data['target'] = data.loan_status == 'Fully Paid'
data.shape
x = data.loan_amnt
term_groups = data.groupby('term')
term_groups['int_rate'].mean()
grade_groups = data.groupby('grade')
grade_groups['int_rate'].mean()
X = pd.get_dummies(data[['term', 'verification_status', 'purpose', 'policy_code', 'loan_amnt', 'funded_amnt', 'funded_amnt_inv', 'int_rate', 'emp_length', 'addr_state']])
X.shape
y = data['target']
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.33, random_state=42)
X_train.shape
clf = RandomForestClassifier(n_estimators=100, max_depth=4, random_state=42)
clf.fit(X_train, y_train)
y_pred = clf.predict(X_test)
y_pred = np.ones(y_test.shape)
accuracy_score(y_test, y_pred) | code |
16127029/cell_11 | [
"text_plain_output_1.png"
] | import matplotlib.pyplot as plt
import pandas as pd # data processing, CSV file I/O (e.g. pd.read_csv)
data = pd.read_csv('../input/loan.csv', low_memory=False)
data = data[(data.loan_status == 'Fully Paid') | (data.loan_status == 'Default')]
data['target'] = data.loan_status == 'Fully Paid'
data.shape
x = data.loan_amnt
term_groups = data.groupby('term')
term_groups['int_rate'].mean()
grade_groups = data.groupby('grade')
grade_groups['int_rate'].mean() | code |
16127029/cell_19 | [
"text_plain_output_1.png"
] | import matplotlib.pyplot as plt
import pandas as pd # data processing, CSV file I/O (e.g. pd.read_csv)
data = pd.read_csv('../input/loan.csv', low_memory=False)
data = data[(data.loan_status == 'Fully Paid') | (data.loan_status == 'Default')]
data['target'] = data.loan_status == 'Fully Paid'
data.shape
x = data.loan_amnt
term_groups = data.groupby('term')
term_groups['int_rate'].mean()
grade_groups = data.groupby('grade')
grade_groups['int_rate'].mean()
X = pd.get_dummies(data[['term', 'verification_status', 'purpose', 'policy_code', 'loan_amnt', 'funded_amnt', 'funded_amnt_inv', 'int_rate', 'emp_length', 'addr_state']])
X.shape | code |
16127029/cell_1 | [
"text_plain_output_1.png"
] | import os
import os
import numpy as np
import pandas as pd
from sklearn.metrics import accuracy_score
from sklearn.model_selection import train_test_split
from sklearn.ensemble import RandomForestClassifier
import matplotlib.pyplot as plt
print(os.listdir('../input')) | code |
16127029/cell_7 | [
"text_plain_output_1.png"
] | import matplotlib.pyplot as plt
import pandas as pd # data processing, CSV file I/O (e.g. pd.read_csv)
data = pd.read_csv('../input/loan.csv', low_memory=False)
data = data[(data.loan_status == 'Fully Paid') | (data.loan_status == 'Default')]
data['target'] = data.loan_status == 'Fully Paid'
data.shape
x = data.loan_amnt
plt.hist(x)
plt.ylabel('No. of loans')
plt.xlabel('Amnt of loan ($)')
plt.show() | code |
16127029/cell_8 | [
"text_plain_output_1.png",
"image_output_1.png"
] | import matplotlib.pyplot as plt
import pandas as pd # data processing, CSV file I/O (e.g. pd.read_csv)
data = pd.read_csv('../input/loan.csv', low_memory=False)
data = data[(data.loan_status == 'Fully Paid') | (data.loan_status == 'Default')]
data['target'] = data.loan_status == 'Fully Paid'
data.shape
x = data.loan_amnt
term_groups = data.groupby('term')
term_groups['int_rate'].mean() | code |
16127029/cell_15 | [
"text_plain_output_1.png"
] | import matplotlib.pyplot as plt
import pandas as pd # data processing, CSV file I/O (e.g. pd.read_csv)
data = pd.read_csv('../input/loan.csv', low_memory=False)
data = data[(data.loan_status == 'Fully Paid') | (data.loan_status == 'Default')]
data['target'] = data.loan_status == 'Fully Paid'
data.shape
x = data.loan_amnt
term_groups = data.groupby('term')
term_groups['int_rate'].mean()
grade_groups = data.groupby('grade')
grade_groups['int_rate'].mean()
total_loaned = grade_groups['funded_amnt'].sum()
total_received = grade_groups['total_pymnt'].sum()
(total_received / total_loaned - 1) * 100 | code |
16127029/cell_16 | [
"text_plain_output_1.png",
"image_output_1.png"
] | import matplotlib.pyplot as plt
import pandas as pd # data processing, CSV file I/O (e.g. pd.read_csv)
data = pd.read_csv('../input/loan.csv', low_memory=False)
data = data[(data.loan_status == 'Fully Paid') | (data.loan_status == 'Default')]
data['target'] = data.loan_status == 'Fully Paid'
data.shape
x = data.loan_amnt
term_groups = data.groupby('term')
term_groups['int_rate'].mean()
grade_groups = data.groupby('grade')
grade_groups['int_rate'].mean()
data['application_type'].value_counts() | code |
16127029/cell_24 | [
"text_plain_output_1.png"
] | from sklearn.ensemble import RandomForestClassifier
from sklearn.metrics import accuracy_score
from sklearn.model_selection import train_test_split
import matplotlib.pyplot as plt
import pandas as pd # data processing, CSV file I/O (e.g. pd.read_csv)
data = pd.read_csv('../input/loan.csv', low_memory=False)
data = data[(data.loan_status == 'Fully Paid') | (data.loan_status == 'Default')]
data['target'] = data.loan_status == 'Fully Paid'
data.shape
x = data.loan_amnt
term_groups = data.groupby('term')
term_groups['int_rate'].mean()
grade_groups = data.groupby('grade')
grade_groups['int_rate'].mean()
X = pd.get_dummies(data[['term', 'verification_status', 'purpose', 'policy_code', 'loan_amnt', 'funded_amnt', 'funded_amnt_inv', 'int_rate', 'emp_length', 'addr_state']])
X.shape
y = data['target']
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.33, random_state=42)
X_train.shape
clf = RandomForestClassifier(n_estimators=100, max_depth=4, random_state=42)
clf.fit(X_train, y_train)
y_pred = clf.predict(X_test)
accuracy_score(y_test, y_pred) | code |
16127029/cell_22 | [
"text_plain_output_1.png"
] | from sklearn.ensemble import RandomForestClassifier
from sklearn.model_selection import train_test_split
import matplotlib.pyplot as plt
import pandas as pd # data processing, CSV file I/O (e.g. pd.read_csv)
data = pd.read_csv('../input/loan.csv', low_memory=False)
data = data[(data.loan_status == 'Fully Paid') | (data.loan_status == 'Default')]
data['target'] = data.loan_status == 'Fully Paid'
data.shape
x = data.loan_amnt
term_groups = data.groupby('term')
term_groups['int_rate'].mean()
grade_groups = data.groupby('grade')
grade_groups['int_rate'].mean()
X = pd.get_dummies(data[['term', 'verification_status', 'purpose', 'policy_code', 'loan_amnt', 'funded_amnt', 'funded_amnt_inv', 'int_rate', 'emp_length', 'addr_state']])
X.shape
y = data['target']
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.33, random_state=42)
X_train.shape
clf = RandomForestClassifier(n_estimators=100, max_depth=4, random_state=42)
clf.fit(X_train, y_train) | code |
16127029/cell_10 | [
"text_plain_output_1.png"
] | import matplotlib.pyplot as plt
import pandas as pd # data processing, CSV file I/O (e.g. pd.read_csv)
data = pd.read_csv('../input/loan.csv', low_memory=False)
data = data[(data.loan_status == 'Fully Paid') | (data.loan_status == 'Default')]
data['target'] = data.loan_status == 'Fully Paid'
data.shape
x = data.loan_amnt
term_groups = data.groupby('term')
term_groups['int_rate'].mean()
data.boxplot('int_rate', by='term') | code |
16127029/cell_12 | [
"image_output_1.png"
] | import matplotlib.pyplot as plt
import pandas as pd # data processing, CSV file I/O (e.g. pd.read_csv)
data = pd.read_csv('../input/loan.csv', low_memory=False)
data = data[(data.loan_status == 'Fully Paid') | (data.loan_status == 'Default')]
data['target'] = data.loan_status == 'Fully Paid'
data.shape
x = data.loan_amnt
term_groups = data.groupby('term')
term_groups['int_rate'].mean()
grade_groups = data.groupby('grade')
grade_groups['int_rate'].mean()
grade_groups['int_rate'].mean().max() | code |
16127029/cell_5 | [
"text_plain_output_1.png"
] | import pandas as pd # data processing, CSV file I/O (e.g. pd.read_csv)
data = pd.read_csv('../input/loan.csv', low_memory=False)
data = data[(data.loan_status == 'Fully Paid') | (data.loan_status == 'Default')]
data['target'] = data.loan_status == 'Fully Paid'
data.shape | code |
34126743/cell_4 | [
"text_html_output_1.png"
] | import pandas as pd
import pandas as pd # data processing, CSV file I/O (e.g. pd.read_csv)
df = pd.read_csv('../input/homicide-reports/database.csv')
df.head() | code |
34126743/cell_3 | [
"application_vnd.jupyter.stderr_output_1.png"
] | import pandas as pd
import pandas as pd # data processing, CSV file I/O (e.g. pd.read_csv)
df = pd.read_csv('../input/homicide-reports/database.csv') | code |
34126743/cell_5 | [
"text_plain_output_1.png",
"image_output_1.png"
] | import matplotlib.pyplot as plt
import pandas as pd
import pandas as pd # data processing, CSV file I/O (e.g. pd.read_csv)
df = pd.read_csv('../input/homicide-reports/database.csv')
plt.style.use('fivethirtyeight')
years = pd.DataFrame(df, columns=['Year'])
count_years = years.stack().value_counts()
homicides = count_years.sort_index(axis=0, ascending=False)
print(homicides.plot(kind='barh', fontsize=10, width=0.5, figsize=(12, 10), title='Отчеты об убийствах в США 1980-2014 ')) | code |
89138279/cell_25 | [
"text_plain_output_1.png"
] | from datasets import Dataset
import pandas as pd
df1 = pd.read_csv('../input/newspaper-articles-csv-ocr/ocr/df_valid.csv')
df1['imgdir'] = '../input/newspaper-articles/articles/validation'
df2 = pd.read_csv('../input/newspaper-articles-csv-ocr/ocr/df_train1.csv')
df2['imgdir'] = '../input/newspaper-articles/articles/training'
df3 = pd.read_csv('../input/newspaper-articles-csv-ocr/ocr/df_train2.csv')
df3['imgdir'] = '../input/newspaper-articles/articles/training'
df4 = pd.read_csv('../input/newspaper-articles-csv-ocr/ocr/df_train3.csv')
df4['imgdir'] = '../input/newspaper-articles/articles/training'
df = pd.concat([df1, df2, df3, df4])
df = df[df.text.apply(lambda x: len(str(x)) > 1)]
df = df.reset_index(drop=True)
df.groupby(['label']).size()
df_meta = df
df_meta = df_meta.drop_duplicates(subset=['image_id'])
df_meta = df_meta.reset_index(drop=True)
df_meta['id'] = df_meta.index
df_meta['image_path'] = df_meta['imgdir'] + '/' + df_meta['image_id']
df_meta = df_meta[['image_id', 'image_path']]
df_meta['split'] = 'train'
ratio_split = int(len(df_meta) * 0.2)
df_meta.loc[df_meta.index[-ratio_split:], 'split'] = 'test'
df_meta
from datasets import Dataset
train_meta = df_meta[df_meta.split == 'train']
train_meta = train_meta.drop(columns=['split'])
train_meta = train_meta.reset_index(drop=True)
dataset_train = Dataset.from_pandas(train_meta)
test_meta = df_meta[df_meta.split == 'test']
test_meta = test_meta.drop(columns=['split'])
test_meta = test_meta.reset_index(drop=True)
dataset_test = Dataset.from_pandas(test_meta)
print(type(dataset_train[0]))
dataset_train[0] | code |
89138279/cell_4 | [
"text_plain_output_1.png"
] | !pip install -U transformers | code |
89138279/cell_23 | [
"text_plain_output_1.png"
] | import pandas as pd
df1 = pd.read_csv('../input/newspaper-articles-csv-ocr/ocr/df_valid.csv')
df1['imgdir'] = '../input/newspaper-articles/articles/validation'
df2 = pd.read_csv('../input/newspaper-articles-csv-ocr/ocr/df_train1.csv')
df2['imgdir'] = '../input/newspaper-articles/articles/training'
df3 = pd.read_csv('../input/newspaper-articles-csv-ocr/ocr/df_train2.csv')
df3['imgdir'] = '../input/newspaper-articles/articles/training'
df4 = pd.read_csv('../input/newspaper-articles-csv-ocr/ocr/df_train3.csv')
df4['imgdir'] = '../input/newspaper-articles/articles/training'
df = pd.concat([df1, df2, df3, df4])
df = df[df.text.apply(lambda x: len(str(x)) > 1)]
df = df.reset_index(drop=True)
df.groupby(['label']).size()
df_meta = df
df_meta = df_meta.drop_duplicates(subset=['image_id'])
df_meta = df_meta.reset_index(drop=True)
df_meta['id'] = df_meta.index
df_meta['image_path'] = df_meta['imgdir'] + '/' + df_meta['image_id']
df_meta = df_meta[['image_id', 'image_path']]
df_meta['split'] = 'train'
print(df_meta.shape)
ratio_split = int(len(df_meta) * 0.2)
df_meta.loc[df_meta.index[-ratio_split:], 'split'] = 'test'
print('size of training_dataset=', len(df_meta[df_meta.split == 'train']), '---- size of test_dataset=', len(df_meta[df_meta.split == 'test']))
df_meta | code |
89138279/cell_30 | [
"text_plain_output_1.png"
] | from PIL import Image
from PIL import Image, ImageDraw, ImageFont
from PIL import Image, ImageDraw, ImageFont
from datasets import Dataset
from datasets import Features, Sequence, ClassLabel, Value, Array2D, Array3D
from transformers import LayoutLMv2Processor
import pandas as pd
df1 = pd.read_csv('../input/newspaper-articles-csv-ocr/ocr/df_valid.csv')
df1['imgdir'] = '../input/newspaper-articles/articles/validation'
df2 = pd.read_csv('../input/newspaper-articles-csv-ocr/ocr/df_train1.csv')
df2['imgdir'] = '../input/newspaper-articles/articles/training'
df3 = pd.read_csv('../input/newspaper-articles-csv-ocr/ocr/df_train2.csv')
df3['imgdir'] = '../input/newspaper-articles/articles/training'
df4 = pd.read_csv('../input/newspaper-articles-csv-ocr/ocr/df_train3.csv')
df4['imgdir'] = '../input/newspaper-articles/articles/training'
df = pd.concat([df1, df2, df3, df4])
df = df[df.text.apply(lambda x: len(str(x)) > 1)]
df = df.reset_index(drop=True)
labels = df['label'].unique()
labels = labels.tolist()
id2label = {v: k for v, k in enumerate(labels)}
label2id = {k: v for v, k in enumerate(labels)}
df.groupby(['label']).size()
label2color = {'title': 'Red', 'text': 'Green'}
def Draw_BBox(image_path, _bbox_labeled: pd.DataFrame):
actual_boxes = []
for idx, row in _bbox_labeled.iterrows():
x0 = row['left']
y0 = row['top']
x1 = row['width'] + row['left']
y1 = row['height'] + row['top']
label = row['label']
conf = row['conf']
color = label2color[label]
actual_box = [x0, y0, x1, y1]
draw = ImageDraw.Draw(image, 'RGB')
draw.rectangle(actual_box, outline=color, width=1)
draw.text((actual_box[0] + 10, actual_box[1] - 10), text=f'{label}', fill=color)
return image
sample_filename = df['image_id'].iloc[0]
sample_dir = df['imgdir'].iloc[0]
from PIL import Image, ImageDraw, ImageFont
image = Image.open(f'{sample_dir}/{sample_filename}')
image = image.convert('RGB').resize((1000, 1000))
sample_image_df = df[df['image_id'] == sample_filename]
sample_image_df = sample_image_df
Draw_BBox(image, sample_image_df)
df_meta = df
df_meta = df_meta.drop_duplicates(subset=['image_id'])
df_meta = df_meta.reset_index(drop=True)
df_meta['id'] = df_meta.index
df_meta['image_path'] = df_meta['imgdir'] + '/' + df_meta['image_id']
df_meta = df_meta[['image_id', 'image_path']]
df_meta['split'] = 'train'
ratio_split = int(len(df_meta) * 0.2)
df_meta.loc[df_meta.index[-ratio_split:], 'split'] = 'test'
df_meta
from datasets import Dataset
train_meta = df_meta[df_meta.split == 'train']
train_meta = train_meta.drop(columns=['split'])
train_meta = train_meta.reset_index(drop=True)
dataset_train = Dataset.from_pandas(train_meta)
test_meta = df_meta[df_meta.split == 'test']
test_meta = test_meta.drop(columns=['split'])
test_meta = test_meta.reset_index(drop=True)
dataset_test = Dataset.from_pandas(test_meta)
def get_words_and_boxes(examples):
image_id = examples['image_id']
record = {}
objs1 = []
objs2 = []
objs3 = []
for index2, row in df.query('image_id == @image_id').iterrows():
class_id = label2id[row['label']]
word = row['text']
bbox_resized = [int(row['left']), int(row['top']), int(row['left']) + int(row['width']), int(row['top']) + int(row['height'])]
obj1 = word
obj2 = bbox_resized
obj3 = class_id
objs1.append(obj1)
objs2.append(obj2)
objs3.append(obj3)
record['words'] = objs1
record['bboxes'] = objs2
record['classes'] = objs3
examples['words'] = [objs1]
examples['bboxes'] = [objs2]
examples['classes'] = [objs3]
return examples
dataset_train_with_bboxs = dataset_train.map(get_words_and_boxes, batched=True, batch_size=1)
dataset_test_with_bboxs = dataset_test.map(get_words_and_boxes, batched=True, batch_size=1)
from PIL import Image, ImageDraw, ImageFont
example = dataset_train_with_bboxs[0]
image = Image.open(example['image_path'])
image = image.convert('RGB')
image
from PIL import Image
from transformers import LayoutLMv2Processor
from datasets import Features, Sequence, ClassLabel, Value, Array2D, Array3D
from transformers import pipeline
processor = LayoutLMv2Processor.from_pretrained('microsoft/layoutlmv2-base-uncased', revision='no_ocr')
features = Features({'image': Array3D(dtype='int64', shape=(3, 224, 224)), 'input_ids': Sequence(feature=Value(dtype='int64')), 'attention_mask': Sequence(Value(dtype='int64')), 'token_type_ids': Sequence(Value(dtype='int64')), 'bbox': Array2D(dtype='int64', shape=(512, 4)), 'labels': Sequence(ClassLabel(names=labels))})
def preprocess_data(examples):
images = [Image.open(path).convert('RGB') for path in examples['image_path']]
words = examples['words']
boxes = examples['bboxes']
word_labels = examples['classes']
encoded_inputs = processor(image, words, boxes=boxes, word_labels=word_labels, padding='max_length', truncation=True)
return encoded_inputs
train_dataset = dataset_train_with_bboxs.map(preprocess_data, batched=True, batch_size=1, remove_columns=dataset_train_with_bboxs.column_names, features=features)
test_dataset = dataset_test_with_bboxs.map(preprocess_data, batched=True, batch_size=1, remove_columns=dataset_test_with_bboxs.column_names, features=features) | code |
89138279/cell_33 | [
"text_plain_output_2.png",
"text_plain_output_1.png"
] | from PIL import Image
from PIL import Image, ImageDraw, ImageFont
from PIL import Image, ImageDraw, ImageFont
from datasets import Dataset
from datasets import Features, Sequence, ClassLabel, Value, Array2D, Array3D
from transformers import LayoutLMv2Processor
import pandas as pd
df1 = pd.read_csv('../input/newspaper-articles-csv-ocr/ocr/df_valid.csv')
df1['imgdir'] = '../input/newspaper-articles/articles/validation'
df2 = pd.read_csv('../input/newspaper-articles-csv-ocr/ocr/df_train1.csv')
df2['imgdir'] = '../input/newspaper-articles/articles/training'
df3 = pd.read_csv('../input/newspaper-articles-csv-ocr/ocr/df_train2.csv')
df3['imgdir'] = '../input/newspaper-articles/articles/training'
df4 = pd.read_csv('../input/newspaper-articles-csv-ocr/ocr/df_train3.csv')
df4['imgdir'] = '../input/newspaper-articles/articles/training'
df = pd.concat([df1, df2, df3, df4])
df = df[df.text.apply(lambda x: len(str(x)) > 1)]
df = df.reset_index(drop=True)
labels = df['label'].unique()
labels = labels.tolist()
id2label = {v: k for v, k in enumerate(labels)}
label2id = {k: v for v, k in enumerate(labels)}
df.groupby(['label']).size()
label2color = {'title': 'Red', 'text': 'Green'}
def Draw_BBox(image_path, _bbox_labeled: pd.DataFrame):
actual_boxes = []
for idx, row in _bbox_labeled.iterrows():
x0 = row['left']
y0 = row['top']
x1 = row['width'] + row['left']
y1 = row['height'] + row['top']
label = row['label']
conf = row['conf']
color = label2color[label]
actual_box = [x0, y0, x1, y1]
draw = ImageDraw.Draw(image, 'RGB')
draw.rectangle(actual_box, outline=color, width=1)
draw.text((actual_box[0] + 10, actual_box[1] - 10), text=f'{label}', fill=color)
return image
sample_filename = df['image_id'].iloc[0]
sample_dir = df['imgdir'].iloc[0]
from PIL import Image, ImageDraw, ImageFont
image = Image.open(f'{sample_dir}/{sample_filename}')
image = image.convert('RGB').resize((1000, 1000))
sample_image_df = df[df['image_id'] == sample_filename]
sample_image_df = sample_image_df
Draw_BBox(image, sample_image_df)
df_meta = df
df_meta = df_meta.drop_duplicates(subset=['image_id'])
df_meta = df_meta.reset_index(drop=True)
df_meta['id'] = df_meta.index
df_meta['image_path'] = df_meta['imgdir'] + '/' + df_meta['image_id']
df_meta = df_meta[['image_id', 'image_path']]
df_meta['split'] = 'train'
ratio_split = int(len(df_meta) * 0.2)
df_meta.loc[df_meta.index[-ratio_split:], 'split'] = 'test'
df_meta
from datasets import Dataset
train_meta = df_meta[df_meta.split == 'train']
train_meta = train_meta.drop(columns=['split'])
train_meta = train_meta.reset_index(drop=True)
dataset_train = Dataset.from_pandas(train_meta)
test_meta = df_meta[df_meta.split == 'test']
test_meta = test_meta.drop(columns=['split'])
test_meta = test_meta.reset_index(drop=True)
dataset_test = Dataset.from_pandas(test_meta)
def get_words_and_boxes(examples):
image_id = examples['image_id']
record = {}
objs1 = []
objs2 = []
objs3 = []
for index2, row in df.query('image_id == @image_id').iterrows():
class_id = label2id[row['label']]
word = row['text']
bbox_resized = [int(row['left']), int(row['top']), int(row['left']) + int(row['width']), int(row['top']) + int(row['height'])]
obj1 = word
obj2 = bbox_resized
obj3 = class_id
objs1.append(obj1)
objs2.append(obj2)
objs3.append(obj3)
record['words'] = objs1
record['bboxes'] = objs2
record['classes'] = objs3
examples['words'] = [objs1]
examples['bboxes'] = [objs2]
examples['classes'] = [objs3]
return examples
dataset_train_with_bboxs = dataset_train.map(get_words_and_boxes, batched=True, batch_size=1)
dataset_test_with_bboxs = dataset_test.map(get_words_and_boxes, batched=True, batch_size=1)
from PIL import Image, ImageDraw, ImageFont
example = dataset_train_with_bboxs[0]
image = Image.open(example['image_path'])
image = image.convert('RGB')
image
from PIL import Image
from transformers import LayoutLMv2Processor
from datasets import Features, Sequence, ClassLabel, Value, Array2D, Array3D
from transformers import pipeline
processor = LayoutLMv2Processor.from_pretrained('microsoft/layoutlmv2-base-uncased', revision='no_ocr')
features = Features({'image': Array3D(dtype='int64', shape=(3, 224, 224)), 'input_ids': Sequence(feature=Value(dtype='int64')), 'attention_mask': Sequence(Value(dtype='int64')), 'token_type_ids': Sequence(Value(dtype='int64')), 'bbox': Array2D(dtype='int64', shape=(512, 4)), 'labels': Sequence(ClassLabel(names=labels))})
def preprocess_data(examples):
images = [Image.open(path).convert('RGB') for path in examples['image_path']]
words = examples['words']
boxes = examples['bboxes']
word_labels = examples['classes']
encoded_inputs = processor(image, words, boxes=boxes, word_labels=word_labels, padding='max_length', truncation=True)
return encoded_inputs
train_dataset = dataset_train_with_bboxs.map(preprocess_data, batched=True, batch_size=1, remove_columns=dataset_train_with_bboxs.column_names, features=features)
test_dataset = dataset_test_with_bboxs.map(preprocess_data, batched=True, batch_size=1, remove_columns=dataset_test_with_bboxs.column_names, features=features)
processor.tokenizer.decode(train_dataset['input_ids'][11]) | code |
89138279/cell_20 | [
"text_html_output_1.png",
"text_plain_output_1.png"
] | from PIL import Image, ImageDraw, ImageFont
import pandas as pd
df1 = pd.read_csv('../input/newspaper-articles-csv-ocr/ocr/df_valid.csv')
df1['imgdir'] = '../input/newspaper-articles/articles/validation'
df2 = pd.read_csv('../input/newspaper-articles-csv-ocr/ocr/df_train1.csv')
df2['imgdir'] = '../input/newspaper-articles/articles/training'
df3 = pd.read_csv('../input/newspaper-articles-csv-ocr/ocr/df_train2.csv')
df3['imgdir'] = '../input/newspaper-articles/articles/training'
df4 = pd.read_csv('../input/newspaper-articles-csv-ocr/ocr/df_train3.csv')
df4['imgdir'] = '../input/newspaper-articles/articles/training'
df = pd.concat([df1, df2, df3, df4])
df = df[df.text.apply(lambda x: len(str(x)) > 1)]
df = df.reset_index(drop=True)
df.groupby(['label']).size()
label2color = {'title': 'Red', 'text': 'Green'}
def Draw_BBox(image_path, _bbox_labeled: pd.DataFrame):
actual_boxes = []
for idx, row in _bbox_labeled.iterrows():
x0 = row['left']
y0 = row['top']
x1 = row['width'] + row['left']
y1 = row['height'] + row['top']
label = row['label']
conf = row['conf']
color = label2color[label]
actual_box = [x0, y0, x1, y1]
draw = ImageDraw.Draw(image, 'RGB')
draw.rectangle(actual_box, outline=color, width=1)
draw.text((actual_box[0] + 10, actual_box[1] - 10), text=f'{label}', fill=color)
return image
sample_filename = df['image_id'].iloc[0]
sample_dir = df['imgdir'].iloc[0]
from PIL import Image, ImageDraw, ImageFont
image = Image.open(f'{sample_dir}/{sample_filename}')
image = image.convert('RGB').resize((1000, 1000))
sample_image_df = df[df['image_id'] == sample_filename]
sample_image_df = sample_image_df
Draw_BBox(image, sample_image_df) | code |
89138279/cell_6 | [
"text_plain_output_1.png"
] | import torch, torchvision
print(torch.__version__, torch.cuda.is_available()) | code |
89138279/cell_29 | [
"text_html_output_1.png",
"text_plain_output_1.png"
] | from PIL import Image, ImageDraw, ImageFont
from PIL import Image, ImageDraw, ImageFont
from datasets import Dataset
import pandas as pd
df1 = pd.read_csv('../input/newspaper-articles-csv-ocr/ocr/df_valid.csv')
df1['imgdir'] = '../input/newspaper-articles/articles/validation'
df2 = pd.read_csv('../input/newspaper-articles-csv-ocr/ocr/df_train1.csv')
df2['imgdir'] = '../input/newspaper-articles/articles/training'
df3 = pd.read_csv('../input/newspaper-articles-csv-ocr/ocr/df_train2.csv')
df3['imgdir'] = '../input/newspaper-articles/articles/training'
df4 = pd.read_csv('../input/newspaper-articles-csv-ocr/ocr/df_train3.csv')
df4['imgdir'] = '../input/newspaper-articles/articles/training'
df = pd.concat([df1, df2, df3, df4])
df = df[df.text.apply(lambda x: len(str(x)) > 1)]
df = df.reset_index(drop=True)
labels = df['label'].unique()
labels = labels.tolist()
id2label = {v: k for v, k in enumerate(labels)}
label2id = {k: v for v, k in enumerate(labels)}
df.groupby(['label']).size()
label2color = {'title': 'Red', 'text': 'Green'}
def Draw_BBox(image_path, _bbox_labeled: pd.DataFrame):
actual_boxes = []
for idx, row in _bbox_labeled.iterrows():
x0 = row['left']
y0 = row['top']
x1 = row['width'] + row['left']
y1 = row['height'] + row['top']
label = row['label']
conf = row['conf']
color = label2color[label]
actual_box = [x0, y0, x1, y1]
draw = ImageDraw.Draw(image, 'RGB')
draw.rectangle(actual_box, outline=color, width=1)
draw.text((actual_box[0] + 10, actual_box[1] - 10), text=f'{label}', fill=color)
return image
sample_filename = df['image_id'].iloc[0]
sample_dir = df['imgdir'].iloc[0]
from PIL import Image, ImageDraw, ImageFont
image = Image.open(f'{sample_dir}/{sample_filename}')
image = image.convert('RGB').resize((1000, 1000))
sample_image_df = df[df['image_id'] == sample_filename]
sample_image_df = sample_image_df
Draw_BBox(image, sample_image_df)
df_meta = df
df_meta = df_meta.drop_duplicates(subset=['image_id'])
df_meta = df_meta.reset_index(drop=True)
df_meta['id'] = df_meta.index
df_meta['image_path'] = df_meta['imgdir'] + '/' + df_meta['image_id']
df_meta = df_meta[['image_id', 'image_path']]
df_meta['split'] = 'train'
ratio_split = int(len(df_meta) * 0.2)
df_meta.loc[df_meta.index[-ratio_split:], 'split'] = 'test'
df_meta
from datasets import Dataset
train_meta = df_meta[df_meta.split == 'train']
train_meta = train_meta.drop(columns=['split'])
train_meta = train_meta.reset_index(drop=True)
dataset_train = Dataset.from_pandas(train_meta)
test_meta = df_meta[df_meta.split == 'test']
test_meta = test_meta.drop(columns=['split'])
test_meta = test_meta.reset_index(drop=True)
dataset_test = Dataset.from_pandas(test_meta)
def get_words_and_boxes(examples):
image_id = examples['image_id']
record = {}
objs1 = []
objs2 = []
objs3 = []
for index2, row in df.query('image_id == @image_id').iterrows():
class_id = label2id[row['label']]
word = row['text']
bbox_resized = [int(row['left']), int(row['top']), int(row['left']) + int(row['width']), int(row['top']) + int(row['height'])]
obj1 = word
obj2 = bbox_resized
obj3 = class_id
objs1.append(obj1)
objs2.append(obj2)
objs3.append(obj3)
record['words'] = objs1
record['bboxes'] = objs2
record['classes'] = objs3
examples['words'] = [objs1]
examples['bboxes'] = [objs2]
examples['classes'] = [objs3]
return examples
dataset_train_with_bboxs = dataset_train.map(get_words_and_boxes, batched=True, batch_size=1)
dataset_test_with_bboxs = dataset_test.map(get_words_and_boxes, batched=True, batch_size=1)
from PIL import Image, ImageDraw, ImageFont
example = dataset_train_with_bboxs[0]
print(example['image_path'])
print('--------------------')
print(example.keys())
print('--------------------')
image = Image.open(example['image_path'])
image = image.convert('RGB')
image | code |
89138279/cell_26 | [
"text_html_output_1.png"
] | from datasets import Dataset
import pandas as pd
df1 = pd.read_csv('../input/newspaper-articles-csv-ocr/ocr/df_valid.csv')
df1['imgdir'] = '../input/newspaper-articles/articles/validation'
df2 = pd.read_csv('../input/newspaper-articles-csv-ocr/ocr/df_train1.csv')
df2['imgdir'] = '../input/newspaper-articles/articles/training'
df3 = pd.read_csv('../input/newspaper-articles-csv-ocr/ocr/df_train2.csv')
df3['imgdir'] = '../input/newspaper-articles/articles/training'
df4 = pd.read_csv('../input/newspaper-articles-csv-ocr/ocr/df_train3.csv')
df4['imgdir'] = '../input/newspaper-articles/articles/training'
df = pd.concat([df1, df2, df3, df4])
df = df[df.text.apply(lambda x: len(str(x)) > 1)]
df = df.reset_index(drop=True)
labels = df['label'].unique()
labels = labels.tolist()
id2label = {v: k for v, k in enumerate(labels)}
label2id = {k: v for v, k in enumerate(labels)}
df.groupby(['label']).size()
df_meta = df
df_meta = df_meta.drop_duplicates(subset=['image_id'])
df_meta = df_meta.reset_index(drop=True)
df_meta['id'] = df_meta.index
df_meta['image_path'] = df_meta['imgdir'] + '/' + df_meta['image_id']
df_meta = df_meta[['image_id', 'image_path']]
df_meta['split'] = 'train'
ratio_split = int(len(df_meta) * 0.2)
df_meta.loc[df_meta.index[-ratio_split:], 'split'] = 'test'
df_meta
from datasets import Dataset
train_meta = df_meta[df_meta.split == 'train']
train_meta = train_meta.drop(columns=['split'])
train_meta = train_meta.reset_index(drop=True)
dataset_train = Dataset.from_pandas(train_meta)
test_meta = df_meta[df_meta.split == 'test']
test_meta = test_meta.drop(columns=['split'])
test_meta = test_meta.reset_index(drop=True)
dataset_test = Dataset.from_pandas(test_meta)
def get_words_and_boxes(examples):
image_id = examples['image_id']
record = {}
objs1 = []
objs2 = []
objs3 = []
for index2, row in df.query('image_id == @image_id').iterrows():
class_id = label2id[row['label']]
word = row['text']
bbox_resized = [int(row['left']), int(row['top']), int(row['left']) + int(row['width']), int(row['top']) + int(row['height'])]
obj1 = word
obj2 = bbox_resized
obj3 = class_id
objs1.append(obj1)
objs2.append(obj2)
objs3.append(obj3)
record['words'] = objs1
record['bboxes'] = objs2
record['classes'] = objs3
examples['words'] = [objs1]
examples['bboxes'] = [objs2]
examples['classes'] = [objs3]
return examples
dataset_train_with_bboxs = dataset_train.map(get_words_and_boxes, batched=True, batch_size=1)
dataset_test_with_bboxs = dataset_test.map(get_words_and_boxes, batched=True, batch_size=1) | code |
89138279/cell_11 | [
"text_plain_output_1.png"
] | import pandas as pd
df1 = pd.read_csv('../input/newspaper-articles-csv-ocr/ocr/df_valid.csv')
df1['imgdir'] = '../input/newspaper-articles/articles/validation'
print(df1.shape)
df2 = pd.read_csv('../input/newspaper-articles-csv-ocr/ocr/df_train1.csv')
df2['imgdir'] = '../input/newspaper-articles/articles/training'
print(df2.shape)
df3 = pd.read_csv('../input/newspaper-articles-csv-ocr/ocr/df_train2.csv')
df3['imgdir'] = '../input/newspaper-articles/articles/training'
print(df3.shape)
df4 = pd.read_csv('../input/newspaper-articles-csv-ocr/ocr/df_train3.csv')
df4['imgdir'] = '../input/newspaper-articles/articles/training'
print(df4.shape)
df = pd.concat([df1, df2, df3, df4])
print(df.shape) | code |
89138279/cell_7 | [
"text_plain_output_1.png"
] | !python -m pip install 'git+https://github.com/facebookresearch/detectron2.git' | code |
89138279/cell_18 | [
"text_plain_output_1.png"
] | import pandas as pd
df1 = pd.read_csv('../input/newspaper-articles-csv-ocr/ocr/df_valid.csv')
df1['imgdir'] = '../input/newspaper-articles/articles/validation'
df2 = pd.read_csv('../input/newspaper-articles-csv-ocr/ocr/df_train1.csv')
df2['imgdir'] = '../input/newspaper-articles/articles/training'
df3 = pd.read_csv('../input/newspaper-articles-csv-ocr/ocr/df_train2.csv')
df3['imgdir'] = '../input/newspaper-articles/articles/training'
df4 = pd.read_csv('../input/newspaper-articles-csv-ocr/ocr/df_train3.csv')
df4['imgdir'] = '../input/newspaper-articles/articles/training'
df = pd.concat([df1, df2, df3, df4])
df = df[df.text.apply(lambda x: len(str(x)) > 1)]
df = df.reset_index(drop=True)
df.groupby(['label']).size()
df | code |
89138279/cell_8 | [
"text_plain_output_1.png"
] | !pip install -U datasets seqeval | code |
89138279/cell_15 | [
"text_plain_output_1.png"
] | import pandas as pd
df1 = pd.read_csv('../input/newspaper-articles-csv-ocr/ocr/df_valid.csv')
df1['imgdir'] = '../input/newspaper-articles/articles/validation'
df2 = pd.read_csv('../input/newspaper-articles-csv-ocr/ocr/df_train1.csv')
df2['imgdir'] = '../input/newspaper-articles/articles/training'
df3 = pd.read_csv('../input/newspaper-articles-csv-ocr/ocr/df_train2.csv')
df3['imgdir'] = '../input/newspaper-articles/articles/training'
df4 = pd.read_csv('../input/newspaper-articles-csv-ocr/ocr/df_train3.csv')
df4['imgdir'] = '../input/newspaper-articles/articles/training'
df = pd.concat([df1, df2, df3, df4])
df = df[df.text.apply(lambda x: len(str(x)) > 1)]
df = df.reset_index(drop=True)
labels = df['label'].unique()
labels = labels.tolist()
print(type(labels))
print(labels) | code |
89138279/cell_16 | [
"text_plain_output_1.png"
] | import pandas as pd
df1 = pd.read_csv('../input/newspaper-articles-csv-ocr/ocr/df_valid.csv')
df1['imgdir'] = '../input/newspaper-articles/articles/validation'
df2 = pd.read_csv('../input/newspaper-articles-csv-ocr/ocr/df_train1.csv')
df2['imgdir'] = '../input/newspaper-articles/articles/training'
df3 = pd.read_csv('../input/newspaper-articles-csv-ocr/ocr/df_train2.csv')
df3['imgdir'] = '../input/newspaper-articles/articles/training'
df4 = pd.read_csv('../input/newspaper-articles-csv-ocr/ocr/df_train3.csv')
df4['imgdir'] = '../input/newspaper-articles/articles/training'
df = pd.concat([df1, df2, df3, df4])
df = df[df.text.apply(lambda x: len(str(x)) > 1)]
df = df.reset_index(drop=True)
labels = df['label'].unique()
labels = labels.tolist()
id2label = {v: k for v, k in enumerate(labels)}
label2id = {k: v for v, k in enumerate(labels)}
print(label2id)
print(id2label) | code |
89138279/cell_17 | [
"text_plain_output_1.png"
] | import pandas as pd
df1 = pd.read_csv('../input/newspaper-articles-csv-ocr/ocr/df_valid.csv')
df1['imgdir'] = '../input/newspaper-articles/articles/validation'
df2 = pd.read_csv('../input/newspaper-articles-csv-ocr/ocr/df_train1.csv')
df2['imgdir'] = '../input/newspaper-articles/articles/training'
df3 = pd.read_csv('../input/newspaper-articles-csv-ocr/ocr/df_train2.csv')
df3['imgdir'] = '../input/newspaper-articles/articles/training'
df4 = pd.read_csv('../input/newspaper-articles-csv-ocr/ocr/df_train3.csv')
df4['imgdir'] = '../input/newspaper-articles/articles/training'
df = pd.concat([df1, df2, df3, df4])
df = df[df.text.apply(lambda x: len(str(x)) > 1)]
df = df.reset_index(drop=True)
df.groupby(['label']).size() | code |
89138279/cell_35 | [
"text_plain_output_1.png",
"image_output_1.png"
] | from PIL import Image
from PIL import Image, ImageDraw, ImageFont
from PIL import Image, ImageDraw, ImageFont
from datasets import Dataset
from datasets import Features, Sequence, ClassLabel, Value, Array2D, Array3D
from transformers import LayoutLMv2Processor
import pandas as pd
df1 = pd.read_csv('../input/newspaper-articles-csv-ocr/ocr/df_valid.csv')
df1['imgdir'] = '../input/newspaper-articles/articles/validation'
df2 = pd.read_csv('../input/newspaper-articles-csv-ocr/ocr/df_train1.csv')
df2['imgdir'] = '../input/newspaper-articles/articles/training'
df3 = pd.read_csv('../input/newspaper-articles-csv-ocr/ocr/df_train2.csv')
df3['imgdir'] = '../input/newspaper-articles/articles/training'
df4 = pd.read_csv('../input/newspaper-articles-csv-ocr/ocr/df_train3.csv')
df4['imgdir'] = '../input/newspaper-articles/articles/training'
df = pd.concat([df1, df2, df3, df4])
df = df[df.text.apply(lambda x: len(str(x)) > 1)]
df = df.reset_index(drop=True)
labels = df['label'].unique()
labels = labels.tolist()
id2label = {v: k for v, k in enumerate(labels)}
label2id = {k: v for v, k in enumerate(labels)}
df.groupby(['label']).size()
label2color = {'title': 'Red', 'text': 'Green'}
def Draw_BBox(image_path, _bbox_labeled: pd.DataFrame):
actual_boxes = []
for idx, row in _bbox_labeled.iterrows():
x0 = row['left']
y0 = row['top']
x1 = row['width'] + row['left']
y1 = row['height'] + row['top']
label = row['label']
conf = row['conf']
color = label2color[label]
actual_box = [x0, y0, x1, y1]
draw = ImageDraw.Draw(image, 'RGB')
draw.rectangle(actual_box, outline=color, width=1)
draw.text((actual_box[0] + 10, actual_box[1] - 10), text=f'{label}', fill=color)
return image
sample_filename = df['image_id'].iloc[0]
sample_dir = df['imgdir'].iloc[0]
from PIL import Image, ImageDraw, ImageFont
image = Image.open(f'{sample_dir}/{sample_filename}')
image = image.convert('RGB').resize((1000, 1000))
sample_image_df = df[df['image_id'] == sample_filename]
sample_image_df = sample_image_df
Draw_BBox(image, sample_image_df)
df_meta = df
df_meta = df_meta.drop_duplicates(subset=['image_id'])
df_meta = df_meta.reset_index(drop=True)
df_meta['id'] = df_meta.index
df_meta['image_path'] = df_meta['imgdir'] + '/' + df_meta['image_id']
df_meta = df_meta[['image_id', 'image_path']]
df_meta['split'] = 'train'
ratio_split = int(len(df_meta) * 0.2)
df_meta.loc[df_meta.index[-ratio_split:], 'split'] = 'test'
df_meta
from datasets import Dataset
train_meta = df_meta[df_meta.split == 'train']
train_meta = train_meta.drop(columns=['split'])
train_meta = train_meta.reset_index(drop=True)
dataset_train = Dataset.from_pandas(train_meta)
test_meta = df_meta[df_meta.split == 'test']
test_meta = test_meta.drop(columns=['split'])
test_meta = test_meta.reset_index(drop=True)
dataset_test = Dataset.from_pandas(test_meta)
def get_words_and_boxes(examples):
image_id = examples['image_id']
record = {}
objs1 = []
objs2 = []
objs3 = []
for index2, row in df.query('image_id == @image_id').iterrows():
class_id = label2id[row['label']]
word = row['text']
bbox_resized = [int(row['left']), int(row['top']), int(row['left']) + int(row['width']), int(row['top']) + int(row['height'])]
obj1 = word
obj2 = bbox_resized
obj3 = class_id
objs1.append(obj1)
objs2.append(obj2)
objs3.append(obj3)
record['words'] = objs1
record['bboxes'] = objs2
record['classes'] = objs3
examples['words'] = [objs1]
examples['bboxes'] = [objs2]
examples['classes'] = [objs3]
return examples
dataset_train_with_bboxs = dataset_train.map(get_words_and_boxes, batched=True, batch_size=1)
dataset_test_with_bboxs = dataset_test.map(get_words_and_boxes, batched=True, batch_size=1)
from PIL import Image, ImageDraw, ImageFont
example = dataset_train_with_bboxs[0]
image = Image.open(example['image_path'])
image = image.convert('RGB')
image
from PIL import Image
from transformers import LayoutLMv2Processor
from datasets import Features, Sequence, ClassLabel, Value, Array2D, Array3D
from transformers import pipeline
processor = LayoutLMv2Processor.from_pretrained('microsoft/layoutlmv2-base-uncased', revision='no_ocr')
features = Features({'image': Array3D(dtype='int64', shape=(3, 224, 224)), 'input_ids': Sequence(feature=Value(dtype='int64')), 'attention_mask': Sequence(Value(dtype='int64')), 'token_type_ids': Sequence(Value(dtype='int64')), 'bbox': Array2D(dtype='int64', shape=(512, 4)), 'labels': Sequence(ClassLabel(names=labels))})
def preprocess_data(examples):
images = [Image.open(path).convert('RGB') for path in examples['image_path']]
words = examples['words']
boxes = examples['bboxes']
word_labels = examples['classes']
encoded_inputs = processor(image, words, boxes=boxes, word_labels=word_labels, padding='max_length', truncation=True)
return encoded_inputs
train_dataset = dataset_train_with_bboxs.map(preprocess_data, batched=True, batch_size=1, remove_columns=dataset_train_with_bboxs.column_names, features=features)
test_dataset = dataset_test_with_bboxs.map(preprocess_data, batched=True, batch_size=1, remove_columns=dataset_test_with_bboxs.column_names, features=features)
train_dataset.set_format(type='torch', device='cuda')
test_dataset.set_format(type='torch', device='cuda')
train_dataset.features.keys() | code |
89138279/cell_31 | [
"text_plain_output_2.png",
"text_plain_output_1.png"
] | from PIL import Image
from PIL import Image, ImageDraw, ImageFont
from PIL import Image, ImageDraw, ImageFont
from datasets import Dataset
from datasets import Features, Sequence, ClassLabel, Value, Array2D, Array3D
from transformers import LayoutLMv2Processor
import pandas as pd
df1 = pd.read_csv('../input/newspaper-articles-csv-ocr/ocr/df_valid.csv')
df1['imgdir'] = '../input/newspaper-articles/articles/validation'
df2 = pd.read_csv('../input/newspaper-articles-csv-ocr/ocr/df_train1.csv')
df2['imgdir'] = '../input/newspaper-articles/articles/training'
df3 = pd.read_csv('../input/newspaper-articles-csv-ocr/ocr/df_train2.csv')
df3['imgdir'] = '../input/newspaper-articles/articles/training'
df4 = pd.read_csv('../input/newspaper-articles-csv-ocr/ocr/df_train3.csv')
df4['imgdir'] = '../input/newspaper-articles/articles/training'
df = pd.concat([df1, df2, df3, df4])
df = df[df.text.apply(lambda x: len(str(x)) > 1)]
df = df.reset_index(drop=True)
labels = df['label'].unique()
labels = labels.tolist()
id2label = {v: k for v, k in enumerate(labels)}
label2id = {k: v for v, k in enumerate(labels)}
df.groupby(['label']).size()
label2color = {'title': 'Red', 'text': 'Green'}
def Draw_BBox(image_path, _bbox_labeled: pd.DataFrame):
actual_boxes = []
for idx, row in _bbox_labeled.iterrows():
x0 = row['left']
y0 = row['top']
x1 = row['width'] + row['left']
y1 = row['height'] + row['top']
label = row['label']
conf = row['conf']
color = label2color[label]
actual_box = [x0, y0, x1, y1]
draw = ImageDraw.Draw(image, 'RGB')
draw.rectangle(actual_box, outline=color, width=1)
draw.text((actual_box[0] + 10, actual_box[1] - 10), text=f'{label}', fill=color)
return image
sample_filename = df['image_id'].iloc[0]
sample_dir = df['imgdir'].iloc[0]
from PIL import Image, ImageDraw, ImageFont
image = Image.open(f'{sample_dir}/{sample_filename}')
image = image.convert('RGB').resize((1000, 1000))
sample_image_df = df[df['image_id'] == sample_filename]
sample_image_df = sample_image_df
Draw_BBox(image, sample_image_df)
df_meta = df
df_meta = df_meta.drop_duplicates(subset=['image_id'])
df_meta = df_meta.reset_index(drop=True)
df_meta['id'] = df_meta.index
df_meta['image_path'] = df_meta['imgdir'] + '/' + df_meta['image_id']
df_meta = df_meta[['image_id', 'image_path']]
df_meta['split'] = 'train'
ratio_split = int(len(df_meta) * 0.2)
df_meta.loc[df_meta.index[-ratio_split:], 'split'] = 'test'
df_meta
from datasets import Dataset
train_meta = df_meta[df_meta.split == 'train']
train_meta = train_meta.drop(columns=['split'])
train_meta = train_meta.reset_index(drop=True)
dataset_train = Dataset.from_pandas(train_meta)
test_meta = df_meta[df_meta.split == 'test']
test_meta = test_meta.drop(columns=['split'])
test_meta = test_meta.reset_index(drop=True)
dataset_test = Dataset.from_pandas(test_meta)
def get_words_and_boxes(examples):
image_id = examples['image_id']
record = {}
objs1 = []
objs2 = []
objs3 = []
for index2, row in df.query('image_id == @image_id').iterrows():
class_id = label2id[row['label']]
word = row['text']
bbox_resized = [int(row['left']), int(row['top']), int(row['left']) + int(row['width']), int(row['top']) + int(row['height'])]
obj1 = word
obj2 = bbox_resized
obj3 = class_id
objs1.append(obj1)
objs2.append(obj2)
objs3.append(obj3)
record['words'] = objs1
record['bboxes'] = objs2
record['classes'] = objs3
examples['words'] = [objs1]
examples['bboxes'] = [objs2]
examples['classes'] = [objs3]
return examples
dataset_train_with_bboxs = dataset_train.map(get_words_and_boxes, batched=True, batch_size=1)
dataset_test_with_bboxs = dataset_test.map(get_words_and_boxes, batched=True, batch_size=1)
from PIL import Image, ImageDraw, ImageFont
example = dataset_train_with_bboxs[0]
image = Image.open(example['image_path'])
image = image.convert('RGB')
image
from PIL import Image
from transformers import LayoutLMv2Processor
from datasets import Features, Sequence, ClassLabel, Value, Array2D, Array3D
from transformers import pipeline
processor = LayoutLMv2Processor.from_pretrained('microsoft/layoutlmv2-base-uncased', revision='no_ocr')
features = Features({'image': Array3D(dtype='int64', shape=(3, 224, 224)), 'input_ids': Sequence(feature=Value(dtype='int64')), 'attention_mask': Sequence(Value(dtype='int64')), 'token_type_ids': Sequence(Value(dtype='int64')), 'bbox': Array2D(dtype='int64', shape=(512, 4)), 'labels': Sequence(ClassLabel(names=labels))})
def preprocess_data(examples):
images = [Image.open(path).convert('RGB') for path in examples['image_path']]
words = examples['words']
boxes = examples['bboxes']
word_labels = examples['classes']
encoded_inputs = processor(image, words, boxes=boxes, word_labels=word_labels, padding='max_length', truncation=True)
return encoded_inputs
train_dataset = dataset_train_with_bboxs.map(preprocess_data, batched=True, batch_size=1, remove_columns=dataset_train_with_bboxs.column_names, features=features)
test_dataset = dataset_test_with_bboxs.map(preprocess_data, batched=True, batch_size=1, remove_columns=dataset_test_with_bboxs.column_names, features=features)
train_dataset[0].keys() | code |
89138279/cell_24 | [
"text_plain_output_1.png"
] | from datasets import Dataset
import pandas as pd
df1 = pd.read_csv('../input/newspaper-articles-csv-ocr/ocr/df_valid.csv')
df1['imgdir'] = '../input/newspaper-articles/articles/validation'
df2 = pd.read_csv('../input/newspaper-articles-csv-ocr/ocr/df_train1.csv')
df2['imgdir'] = '../input/newspaper-articles/articles/training'
df3 = pd.read_csv('../input/newspaper-articles-csv-ocr/ocr/df_train2.csv')
df3['imgdir'] = '../input/newspaper-articles/articles/training'
df4 = pd.read_csv('../input/newspaper-articles-csv-ocr/ocr/df_train3.csv')
df4['imgdir'] = '../input/newspaper-articles/articles/training'
df = pd.concat([df1, df2, df3, df4])
df = df[df.text.apply(lambda x: len(str(x)) > 1)]
df = df.reset_index(drop=True)
df.groupby(['label']).size()
df_meta = df
df_meta = df_meta.drop_duplicates(subset=['image_id'])
df_meta = df_meta.reset_index(drop=True)
df_meta['id'] = df_meta.index
df_meta['image_path'] = df_meta['imgdir'] + '/' + df_meta['image_id']
df_meta = df_meta[['image_id', 'image_path']]
df_meta['split'] = 'train'
ratio_split = int(len(df_meta) * 0.2)
df_meta.loc[df_meta.index[-ratio_split:], 'split'] = 'test'
df_meta
from datasets import Dataset
train_meta = df_meta[df_meta.split == 'train']
train_meta = train_meta.drop(columns=['split'])
train_meta = train_meta.reset_index(drop=True)
dataset_train = Dataset.from_pandas(train_meta)
print(dataset_train)
test_meta = df_meta[df_meta.split == 'test']
test_meta = test_meta.drop(columns=['split'])
test_meta = test_meta.reset_index(drop=True)
dataset_test = Dataset.from_pandas(test_meta)
print(dataset_test) | code |
89138279/cell_14 | [
"text_plain_output_1.png"
] | import pandas as pd
df1 = pd.read_csv('../input/newspaper-articles-csv-ocr/ocr/df_valid.csv')
df1['imgdir'] = '../input/newspaper-articles/articles/validation'
df2 = pd.read_csv('../input/newspaper-articles-csv-ocr/ocr/df_train1.csv')
df2['imgdir'] = '../input/newspaper-articles/articles/training'
df3 = pd.read_csv('../input/newspaper-articles-csv-ocr/ocr/df_train2.csv')
df3['imgdir'] = '../input/newspaper-articles/articles/training'
df4 = pd.read_csv('../input/newspaper-articles-csv-ocr/ocr/df_train3.csv')
df4['imgdir'] = '../input/newspaper-articles/articles/training'
df = pd.concat([df1, df2, df3, df4])
df = df[df.text.apply(lambda x: len(str(x)) > 1)]
df = df.reset_index(drop=True)
print(df.shape)
df.head() | code |
89138279/cell_27 | [
"image_output_1.png"
] | from datasets import Dataset
import pandas as pd
df1 = pd.read_csv('../input/newspaper-articles-csv-ocr/ocr/df_valid.csv')
df1['imgdir'] = '../input/newspaper-articles/articles/validation'
df2 = pd.read_csv('../input/newspaper-articles-csv-ocr/ocr/df_train1.csv')
df2['imgdir'] = '../input/newspaper-articles/articles/training'
df3 = pd.read_csv('../input/newspaper-articles-csv-ocr/ocr/df_train2.csv')
df3['imgdir'] = '../input/newspaper-articles/articles/training'
df4 = pd.read_csv('../input/newspaper-articles-csv-ocr/ocr/df_train3.csv')
df4['imgdir'] = '../input/newspaper-articles/articles/training'
df = pd.concat([df1, df2, df3, df4])
df = df[df.text.apply(lambda x: len(str(x)) > 1)]
df = df.reset_index(drop=True)
labels = df['label'].unique()
labels = labels.tolist()
id2label = {v: k for v, k in enumerate(labels)}
label2id = {k: v for v, k in enumerate(labels)}
df.groupby(['label']).size()
df_meta = df
df_meta = df_meta.drop_duplicates(subset=['image_id'])
df_meta = df_meta.reset_index(drop=True)
df_meta['id'] = df_meta.index
df_meta['image_path'] = df_meta['imgdir'] + '/' + df_meta['image_id']
df_meta = df_meta[['image_id', 'image_path']]
df_meta['split'] = 'train'
ratio_split = int(len(df_meta) * 0.2)
df_meta.loc[df_meta.index[-ratio_split:], 'split'] = 'test'
df_meta
from datasets import Dataset
train_meta = df_meta[df_meta.split == 'train']
train_meta = train_meta.drop(columns=['split'])
train_meta = train_meta.reset_index(drop=True)
dataset_train = Dataset.from_pandas(train_meta)
test_meta = df_meta[df_meta.split == 'test']
test_meta = test_meta.drop(columns=['split'])
test_meta = test_meta.reset_index(drop=True)
dataset_test = Dataset.from_pandas(test_meta)
def get_words_and_boxes(examples):
image_id = examples['image_id']
record = {}
objs1 = []
objs2 = []
objs3 = []
for index2, row in df.query('image_id == @image_id').iterrows():
class_id = label2id[row['label']]
word = row['text']
bbox_resized = [int(row['left']), int(row['top']), int(row['left']) + int(row['width']), int(row['top']) + int(row['height'])]
obj1 = word
obj2 = bbox_resized
obj3 = class_id
objs1.append(obj1)
objs2.append(obj2)
objs3.append(obj3)
record['words'] = objs1
record['bboxes'] = objs2
record['classes'] = objs3
examples['words'] = [objs1]
examples['bboxes'] = [objs2]
examples['classes'] = [objs3]
return examples
dataset_train_with_bboxs = dataset_train.map(get_words_and_boxes, batched=True, batch_size=1)
dataset_test_with_bboxs = dataset_test.map(get_words_and_boxes, batched=True, batch_size=1)
print(type(dataset_train_with_bboxs))
print('##############################')
print(type(dataset_train_with_bboxs[0]))
print('##############################')
dataset_train_with_bboxs | code |
89138279/cell_5 | [
"text_plain_output_1.png"
] | pip install -U transformers tokenizers | code |
128034962/cell_9 | [
"image_output_11.png",
"image_output_24.png",
"image_output_25.png",
"text_plain_output_5.png",
"text_plain_output_15.png",
"image_output_17.png",
"text_plain_output_9.png",
"image_output_14.png",
"text_plain_output_20.png",
"image_output_23.png",
"text_plain_output_4.png",
"text_plain_output_13.png",
"image_output_13.png",
"image_output_5.png",
"text_plain_output_14.png",
"image_output_18.png",
"image_output_21.png",
"text_plain_output_27.png",
"text_plain_output_10.png",
"text_plain_output_6.png",
"image_output_7.png",
"text_plain_output_24.png",
"text_plain_output_21.png",
"text_plain_output_25.png",
"image_output_20.png",
"text_plain_output_18.png",
"text_plain_output_3.png",
"image_output_4.png",
"text_plain_output_22.png",
"text_plain_output_7.png",
"image_output_8.png",
"text_plain_output_16.png",
"image_output_16.png",
"text_plain_output_8.png",
"text_plain_output_26.png",
"image_output_27.png",
"image_output_6.png",
"text_plain_output_23.png",
"image_output_12.png",
"image_output_22.png",
"text_plain_output_2.png",
"text_plain_output_1.png",
"image_output_3.png",
"text_plain_output_19.png",
"image_output_2.png",
"image_output_1.png",
"image_output_10.png",
"text_plain_output_17.png",
"text_plain_output_11.png",
"text_plain_output_12.png",
"image_output_15.png",
"image_output_9.png",
"image_output_19.png",
"image_output_26.png"
] | from tensorflow.keras.layers import Conv2D, MaxPooling2D, Flatten, Dense, Dropout
from tensorflow.keras.preprocessing.image import ImageDataGenerator
import matplotlib.pyplot as plt
import tensorflow as tf
import os
import numpy as np
import tensorflow as tf
from tensorflow.keras.preprocessing.image import ImageDataGenerator
from tensorflow.keras.layers import Conv2D, MaxPooling2D, Flatten, Dense, Dropout
from sklearn.model_selection import train_test_split
import matplotlib.pyplot as plt
data_dir = 'images/'
img_width, img_height = (128, 128)
batch_size = 32
train_datagen = ImageDataGenerator(rescale=1.0 / 255, validation_split=0.2)
train_generator = train_datagen.flow_from_directory(data_dir, target_size=(img_width, img_height), batch_size=batch_size, class_mode='categorical', color_mode='grayscale', subset='training')
validation_generator = train_datagen.flow_from_directory(data_dir, target_size=(img_width, img_height), batch_size=batch_size, class_mode='categorical', color_mode='grayscale', subset='validation')
def create_alexnet(input_shape, num_classes, activation):
model = tf.keras.Sequential([Conv2D(96, (11, 11), activation=activation, input_shape=input_shape, strides=(4, 4), padding='same'), MaxPooling2D(pool_size=(3, 3), strides=(2, 2)), Conv2D(256, (5, 5), activation=activation, strides=(1, 1), padding='same'), MaxPooling2D(pool_size=(3, 3), strides=(2, 2)), Conv2D(384, (3, 3), activation=activation, strides=(1, 1), padding='same'), Conv2D(384, (3, 3), activation=activation, strides=(1, 1), padding='same'), Conv2D(256, (3, 3), activation=activation, strides=(1, 1), padding='same'), MaxPooling2D(pool_size=(3, 3), strides=(2, 2)), Flatten(), Dense(4096, activation=activation), Dropout(0.5), Dense(4096, activation=activation), Dropout(0.5), Dense(num_classes, activation='softmax')])
return model
'\n# Define learning rate adjustment schemes\ndef lr_schedule_1(epoch, lr):\n if epoch < 20:\n return lr\n elif epoch < 40:\n return lr * 0.1\n else:\n return lr * 0.01\n\ndef lr_schedule_2(epoch, lr):\n return lr * (1 / (1 + 0.1 * epoch))\n'
num_classes = 8
input_shape = (img_width, img_height, 1)
activations = ['relu', 'sigmoid', 'tanh']
optimizers = [tf.keras.optimizers.legacy.Adam(learning_rate=0.001), tf.keras.optimizers.legacy.Adam(learning_rate=0.0001), tf.keras.optimizers.legacy.Adam(learning_rate=1e-05), tf.keras.optimizers.legacy.SGD(learning_rate=0.001), tf.keras.optimizers.legacy.SGD(learning_rate=0.0001), tf.keras.optimizers.legacy.SGD(learning_rate=1e-05), tf.keras.optimizers.legacy.RMSprop(learning_rate=0.001), tf.keras.optimizers.legacy.RMSprop(learning_rate=0.0001), tf.keras.optimizers.legacy.RMSprop(learning_rate=1e-05)]
adams = optimizers[:3]
sgds = optimizers[3:6]
rmsprops = optimizers[6:]
for activation in activations:
for optimizer in adams:
initial_lr = round(float(optimizer.learning_rate), 5)
optimizer._name = 'Adam'
model = create_alexnet(input_shape, num_classes, activation)
model.compile(optimizer=optimizer, loss='categorical_crossentropy', metrics=['accuracy'])
history = model.fit(train_generator, epochs=50, validation_data=validation_generator)
for activation in activations:
for optimizer in sgds:
initial_lr = round(float(optimizer.learning_rate), 5)
optimizer._name = 'SGD'
print(f'Training with {activation} activation and optimizer {optimizer._name} with {initial_lr}')
model = create_alexnet(input_shape, num_classes, activation)
model.compile(optimizer=optimizer, loss='categorical_crossentropy', metrics=['accuracy'])
history = model.fit(train_generator, epochs=50, validation_data=validation_generator)
plt.figure(figsize=(12, 6))
plt.subplot(1, 2, 1)
plt.plot(history.history['loss'], label='Train Loss')
plt.plot(history.history['val_loss'], label='Validation Loss')
plt.xlabel('Epochs')
plt.ylabel('Loss')
plt.title(f'Loss per epoch with {activation} and optimizer {optimizer._name} with {initial_lr}')
plt.legend()
plt.subplot(1, 2, 2)
plt.plot(history.history['accuracy'], label='Train Accuracy')
plt.plot(history.history['val_accuracy'], label='Validation Accuracy')
plt.xlabel('Epochs')
plt.ylabel('Accuracy')
plt.title(f'Accuracy per epoch with {activation} and optimizer {optimizer._name} with {initial_lr}')
plt.legend()
plt.show() | code |
128034962/cell_6 | [
"text_plain_output_5.png",
"text_plain_output_9.png",
"text_plain_output_4.png",
"image_output_5.png",
"text_plain_output_6.png",
"image_output_7.png",
"text_plain_output_3.png",
"image_output_4.png",
"text_plain_output_7.png",
"image_output_8.png",
"text_plain_output_8.png",
"image_output_6.png",
"text_plain_output_2.png",
"text_plain_output_1.png",
"image_output_3.png",
"image_output_2.png",
"image_output_1.png",
"image_output_9.png"
] | code |
|
128034962/cell_8 | [
"text_plain_output_5.png",
"text_plain_output_9.png",
"text_plain_output_4.png",
"image_output_5.png",
"text_plain_output_6.png",
"image_output_7.png",
"text_plain_output_3.png",
"image_output_4.png",
"text_plain_output_7.png",
"image_output_8.png",
"text_plain_output_8.png",
"image_output_6.png",
"text_plain_output_2.png",
"text_plain_output_1.png",
"image_output_3.png",
"image_output_2.png",
"image_output_1.png",
"image_output_9.png"
] | from tensorflow.keras.layers import Conv2D, MaxPooling2D, Flatten, Dense, Dropout
from tensorflow.keras.preprocessing.image import ImageDataGenerator
import matplotlib.pyplot as plt
import tensorflow as tf
import os
import numpy as np
import tensorflow as tf
from tensorflow.keras.preprocessing.image import ImageDataGenerator
from tensorflow.keras.layers import Conv2D, MaxPooling2D, Flatten, Dense, Dropout
from sklearn.model_selection import train_test_split
import matplotlib.pyplot as plt
data_dir = 'images/'
img_width, img_height = (128, 128)
batch_size = 32
train_datagen = ImageDataGenerator(rescale=1.0 / 255, validation_split=0.2)
train_generator = train_datagen.flow_from_directory(data_dir, target_size=(img_width, img_height), batch_size=batch_size, class_mode='categorical', color_mode='grayscale', subset='training')
validation_generator = train_datagen.flow_from_directory(data_dir, target_size=(img_width, img_height), batch_size=batch_size, class_mode='categorical', color_mode='grayscale', subset='validation')
def create_alexnet(input_shape, num_classes, activation):
model = tf.keras.Sequential([Conv2D(96, (11, 11), activation=activation, input_shape=input_shape, strides=(4, 4), padding='same'), MaxPooling2D(pool_size=(3, 3), strides=(2, 2)), Conv2D(256, (5, 5), activation=activation, strides=(1, 1), padding='same'), MaxPooling2D(pool_size=(3, 3), strides=(2, 2)), Conv2D(384, (3, 3), activation=activation, strides=(1, 1), padding='same'), Conv2D(384, (3, 3), activation=activation, strides=(1, 1), padding='same'), Conv2D(256, (3, 3), activation=activation, strides=(1, 1), padding='same'), MaxPooling2D(pool_size=(3, 3), strides=(2, 2)), Flatten(), Dense(4096, activation=activation), Dropout(0.5), Dense(4096, activation=activation), Dropout(0.5), Dense(num_classes, activation='softmax')])
return model
'\n# Define learning rate adjustment schemes\ndef lr_schedule_1(epoch, lr):\n if epoch < 20:\n return lr\n elif epoch < 40:\n return lr * 0.1\n else:\n return lr * 0.01\ndef lr_schedule_2(epoch, lr):\n return lr * (1 / (1 + 0.1 * epoch))\n'
num_classes = 8
input_shape = (img_width, img_height, 1)
activations = ['relu', 'sigmoid', 'tanh']
optimizers = [tf.keras.optimizers.legacy.Adam(learning_rate=0.001), tf.keras.optimizers.legacy.Adam(learning_rate=0.0001), tf.keras.optimizers.legacy.Adam(learning_rate=1e-05), tf.keras.optimizers.legacy.SGD(learning_rate=0.001), tf.keras.optimizers.legacy.SGD(learning_rate=0.0001), tf.keras.optimizers.legacy.SGD(learning_rate=1e-05), tf.keras.optimizers.legacy.RMSprop(learning_rate=0.001), tf.keras.optimizers.legacy.RMSprop(learning_rate=0.0001), tf.keras.optimizers.legacy.RMSprop(learning_rate=1e-05)]
adams = optimizers[:3]
sgds = optimizers[3:6]
rmsprops = optimizers[6:]
for activation in activations:
for optimizer in adams:
initial_lr = round(float(optimizer.learning_rate), 5)
optimizer._name = 'Adam'
print(f'Training with {activation} activation and optimizer {optimizer._name} with {initial_lr}')
model = create_alexnet(input_shape, num_classes, activation)
model.compile(optimizer=optimizer, loss='categorical_crossentropy', metrics=['accuracy'])
history = model.fit(train_generator, epochs=50, validation_data=validation_generator)
plt.figure(figsize=(12, 6))
plt.subplot(1, 2, 1)
plt.plot(history.history['loss'], label='Train Loss')
plt.plot(history.history['val_loss'], label='Validation Loss')
plt.xlabel('Epochs')
plt.ylabel('Loss')
plt.title(f'Loss per epoch with {activation} and optimizer {optimizer._name} with {initial_lr}')
plt.legend()
plt.subplot(1, 2, 2)
plt.plot(history.history['accuracy'], label='Train Accuracy')
plt.plot(history.history['val_accuracy'], label='Validation Accuracy')
plt.xlabel('Epochs')
plt.ylabel('Accuracy')
plt.title(f'Accuracy per epoch with {activation} and optimizer {optimizer._name} with {initial_lr}')
plt.legend()
plt.show() | code |
128034962/cell_10 | [
"image_output_11.png",
"image_output_24.png",
"text_plain_output_5.png",
"image_output_17.png",
"image_output_14.png",
"image_output_23.png",
"text_plain_output_4.png",
"image_output_13.png",
"image_output_5.png",
"image_output_18.png",
"image_output_21.png",
"text_plain_output_6.png",
"image_output_7.png",
"image_output_20.png",
"text_plain_output_3.png",
"image_output_4.png",
"text_plain_output_7.png",
"image_output_8.png",
"image_output_16.png",
"text_plain_output_8.png",
"image_output_6.png",
"image_output_12.png",
"image_output_22.png",
"text_plain_output_2.png",
"text_plain_output_1.png",
"image_output_3.png",
"image_output_2.png",
"image_output_1.png",
"image_output_10.png",
"image_output_15.png",
"image_output_9.png",
"image_output_19.png"
] | from tensorflow.keras.layers import Conv2D, MaxPooling2D, Flatten, Dense, Dropout
from tensorflow.keras.preprocessing.image import ImageDataGenerator
import matplotlib.pyplot as plt
import tensorflow as tf
import os
import numpy as np
import tensorflow as tf
from tensorflow.keras.preprocessing.image import ImageDataGenerator
from tensorflow.keras.layers import Conv2D, MaxPooling2D, Flatten, Dense, Dropout
from sklearn.model_selection import train_test_split
import matplotlib.pyplot as plt
data_dir = 'images/'
img_width, img_height = (128, 128)
batch_size = 32
train_datagen = ImageDataGenerator(rescale=1.0 / 255, validation_split=0.2)
train_generator = train_datagen.flow_from_directory(data_dir, target_size=(img_width, img_height), batch_size=batch_size, class_mode='categorical', color_mode='grayscale', subset='training')
validation_generator = train_datagen.flow_from_directory(data_dir, target_size=(img_width, img_height), batch_size=batch_size, class_mode='categorical', color_mode='grayscale', subset='validation')
def create_alexnet(input_shape, num_classes, activation):
model = tf.keras.Sequential([Conv2D(96, (11, 11), activation=activation, input_shape=input_shape, strides=(4, 4), padding='same'), MaxPooling2D(pool_size=(3, 3), strides=(2, 2)), Conv2D(256, (5, 5), activation=activation, strides=(1, 1), padding='same'), MaxPooling2D(pool_size=(3, 3), strides=(2, 2)), Conv2D(384, (3, 3), activation=activation, strides=(1, 1), padding='same'), Conv2D(384, (3, 3), activation=activation, strides=(1, 1), padding='same'), Conv2D(256, (3, 3), activation=activation, strides=(1, 1), padding='same'), MaxPooling2D(pool_size=(3, 3), strides=(2, 2)), Flatten(), Dense(4096, activation=activation), Dropout(0.5), Dense(4096, activation=activation), Dropout(0.5), Dense(num_classes, activation='softmax')])
return model
'\n# Define learning rate adjustment schemes\ndef lr_schedule_1(epoch, lr):\n if epoch < 20:\n return lr\n elif epoch < 40:\n return lr * 0.1\n else:\n return lr * 0.01\n\ndef lr_schedule_2(epoch, lr):\n return lr * (1 / (1 + 0.1 * epoch))\n'
num_classes = 8
input_shape = (img_width, img_height, 1)
activations = ['relu', 'sigmoid', 'tanh']
optimizers = [tf.keras.optimizers.legacy.Adam(learning_rate=0.001), tf.keras.optimizers.legacy.Adam(learning_rate=0.0001), tf.keras.optimizers.legacy.Adam(learning_rate=1e-05), tf.keras.optimizers.legacy.SGD(learning_rate=0.001), tf.keras.optimizers.legacy.SGD(learning_rate=0.0001), tf.keras.optimizers.legacy.SGD(learning_rate=1e-05), tf.keras.optimizers.legacy.RMSprop(learning_rate=0.001), tf.keras.optimizers.legacy.RMSprop(learning_rate=0.0001), tf.keras.optimizers.legacy.RMSprop(learning_rate=1e-05)]
adams = optimizers[:3]
sgds = optimizers[3:6]
rmsprops = optimizers[6:]
for activation in activations:
for optimizer in adams:
initial_lr = round(float(optimizer.learning_rate), 5)
optimizer._name = 'Adam'
model = create_alexnet(input_shape, num_classes, activation)
model.compile(optimizer=optimizer, loss='categorical_crossentropy', metrics=['accuracy'])
history = model.fit(train_generator, epochs=50, validation_data=validation_generator)
for activation in activations:
for optimizer in sgds:
initial_lr = round(float(optimizer.learning_rate), 5)
optimizer._name = 'SGD'
model = create_alexnet(input_shape, num_classes, activation)
model.compile(optimizer=optimizer, loss='categorical_crossentropy', metrics=['accuracy'])
history = model.fit(train_generator, epochs=50, validation_data=validation_generator)
for activation in activations:
for optimizer in rmsprops:
initial_lr = round(float(optimizer.learning_rate), 5)
optimizer._name = 'RMSprop'
print(f'Training with {activation} activation and optimizer {optimizer._name} with {initial_lr}')
model = create_alexnet(input_shape, num_classes, activation)
model.compile(optimizer=optimizer, loss='categorical_crossentropy', metrics=['accuracy'])
history = model.fit(train_generator, epochs=50, validation_data=validation_generator)
plt.figure(figsize=(12, 6))
plt.subplot(1, 2, 1)
plt.plot(history.history['loss'], label='Train Loss')
plt.plot(history.history['val_loss'], label='Validation Loss')
plt.xlabel('Epochs')
plt.ylabel('Loss')
plt.title(f'Loss per epoch with {activation} and optimizer {optimizer._name} with {initial_lr}')
plt.legend()
plt.subplot(1, 2, 2)
plt.plot(history.history['accuracy'], label='Train Accuracy')
plt.plot(history.history['val_accuracy'], label='Validation Accuracy')
plt.xlabel('Epochs')
plt.ylabel('Accuracy')
plt.title(f'Accuracy per epoch with {activation} and optimizer {optimizer._name} with {initial_lr}')
plt.legend()
plt.show() | code |
128034962/cell_5 | [
"text_plain_output_5.png",
"text_plain_output_9.png",
"text_plain_output_4.png",
"image_output_5.png",
"text_plain_output_6.png",
"image_output_7.png",
"text_plain_output_3.png",
"image_output_4.png",
"text_plain_output_7.png",
"image_output_8.png",
"text_plain_output_8.png",
"image_output_6.png",
"text_plain_output_2.png",
"text_plain_output_1.png",
"image_output_3.png",
"image_output_2.png",
"image_output_1.png",
"image_output_9.png"
] | from IPython.display import Image, display
from random import randint, random
from sklearn.model_selection import train_test_split
import cv2
import numpy as np
import os
import os
import random
import os
import cv2
import numpy as np
from random import randint, random
from sklearn.model_selection import train_test_split
def random_color():
return randint(128, 255)
def random_polygon(image, vertices):
center = (randint(10, 118), randint(10, 118))
radius = randint(5, 60)
color = random_color()
angle = 2 * np.pi / vertices
pts = []
for i in range(vertices):
x = int(center[0] + radius * np.cos(i * angle))
y = int(center[1] + radius * np.sin(i * angle))
pts.append((x, y))
pts = np.array(pts, np.int32)
cv2.fillPoly(image, [pts], color)
def random_oval(image):
thickness = -1
center = (randint(10, 118), randint(10, 118))
axes = (randint(5, 60), randint(5, 60))
angle = randint(0, 360)
color = random_color()
cv2.ellipse(image, center, axes, angle, 0, 360, color, thickness)
def random_rectangle(image):
thickness = -1
pt1 = (randint(5, 123), randint(5, 123))
pt2 = (randint(5, 123), randint(5, 123))
color = random_color()
cv2.rectangle(image, pt1, pt2, color, thickness)
def random_triangle(image):
thickness = -1
pts = np.array([[randint(5, 123), randint(5, 123)], [randint(5, 123), randint(5, 123)], [randint(5, 123), randint(5, 123)]], np.int32)
color = random_color()
cv2.fillPoly(image, [pts], color)
def random_star(image, vertices):
center = (randint(10, 118), randint(10, 118))
inner_radius = randint(5, 30)
outer_radius = randint(inner_radius + 1, 60)
color = random_color()
angle = 2 * np.pi / vertices
inner_pts = []
outer_pts = []
for i in range(vertices):
inner_x = int(center[0] + inner_radius * np.cos(i * angle))
inner_y = int(center[1] + inner_radius * np.sin(i * angle))
outer_x = int(center[0] + outer_radius * np.cos(i * angle + angle / 2))
outer_y = int(center[1] + outer_radius * np.sin(i * angle + angle / 2))
inner_pts.append((inner_x, inner_y))
outer_pts.append((outer_x, outer_y))
pts = [pt for pair in zip(inner_pts, outer_pts) for pt in pair]
pts = np.array(pts, np.int32)
cv2.fillPoly(image, [pts], color)
def add_salt_pepper_noise(image, prob):
row, col = image.shape
sp_image = np.copy(image)
num_salt = np.ceil(prob * image.size)
coords = [np.random.randint(0, i - 1, int(num_salt)) for i in image.shape]
sp_image[tuple(coords)] = 255
num_pepper = np.ceil(prob * image.size)
coords = [np.random.randint(0, i - 1, int(num_pepper)) for i in image.shape]
sp_image[tuple(coords)] = 0
return sp_image
def generate_images(num_images, shape_func, shape_name):
for i in range(num_images):
image = np.zeros((128, 128), dtype=np.uint8)
shape_func(image)
image_with_noise = add_salt_pepper_noise(image, 0.01)
cv2.imwrite(f'images/{shape_name}/{shape_name}_{i:04d}.png', image_with_noise)
def create_directory_structure():
os.makedirs('images', exist_ok=True)
for shape in ['oval', 'rectangle', 'triangle', 'poly5', 'poly6', 'poly7', 'star5', 'star8']:
os.makedirs(f'images/{shape}', exist_ok=True)
def main():
create_directory_structure()
np.random.seed(42)
num_images = 1000
generate_images(num_images, random_oval, 'oval')
generate_images(num_images, random_rectangle, 'rectangle')
generate_images(num_images, random_triangle, 'triangle')
for sides in [5, 6, 7]:
generate_images(num_images, lambda img: random_polygon(img, sides), f'poly{sides}')
generate_images(num_images, lambda img: random_star(img, 5), 'star5')
generate_images(num_images, lambda img: random_star(img, 8), 'star8')
X = []
y = []
for i, shape in enumerate(['oval', 'rectangle', 'triangle', 'poly5', 'poly6', 'poly7', 'star5', 'star8']):
for j in range(num_images):
img = cv2.imread(f'images/{shape}/{shape}_{j:04d}.png')
X.append(img)
y.append(i)
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)
np.save('X_train.npy', X_train)
np.save('X_test.npy', X_test)
np.save('y_train.npy', y_train)
np.save('y_test.npy', y_test)
main()
import os
from IPython.display import Image, display
import random
shapes = ['oval', 'poly5', 'poly6', 'poly7', 'rectangle', 'star5', 'star8', 'triangle']
for shape in shapes:
print(f'\nSample images for {shape}:')
files = os.listdir(f'./images/{shape}')
for i in range(3):
file = random.choice(files)
display(Image(f'./images/{shape}/{file}')) | code |
333930/cell_1 | [
"application_vnd.jupyter.stderr_output_1.png"
] | library(ggplot2)
library(readr)
library(dplyr)
library(tidyr)
library(DT)
system('ls ../input') | code |
121151894/cell_4 | [
"text_html_output_1.png"
] | import pandas as pd
df = pd.read_csv('/kaggle/input/reliance-nse-stock-data/RILO - Copy.csv')
df.sample(3)
x = df[['Prev Close', 'Open Price', 'High Price', 'Low Price', 'Close Price', 'Deliverable Qty', '% Dly Qt to Traded Qty']]
x | code |
121151894/cell_6 | [
"text_plain_output_1.png"
] | import pandas as pd
df = pd.read_csv('/kaggle/input/reliance-nse-stock-data/RILO - Copy.csv')
df.sample(3)
y = df.iloc[:, 9]
df.shape | code |
121151894/cell_29 | [
"text_plain_output_1.png"
] | from sklearn.linear_model import LinearRegression,LogisticRegression
from sklearn.model_selection import train_test_split,cross_val_score
lr = LinearRegression()
lr.fit(x_train, y_train)
lr.predict(x_test)
y_pred = lr.predict(x_test)
cross_val_score(lr, x_train, y_train, scoring='r2').mean() | code |
121151894/cell_26 | [
"text_plain_output_1.png"
] | from sklearn.linear_model import LinearRegression,LogisticRegression
from sklearn.preprocessing import StandardScaler,MinMaxScaler
ss = StandardScaler()
x_train_ss = ss.fit_transform(x_train)
x_test_ss = ss.fit_transform(x_test)
lr2 = LinearRegression()
lr2.fit(x_train_ss, y_train) | code |
121151894/cell_11 | [
"text_html_output_1.png"
] | from sklearn.linear_model import LinearRegression,LogisticRegression
lr = LinearRegression()
lr.fit(x_train, y_train) | code |
121151894/cell_7 | [
"text_plain_output_1.png"
] | import pandas as pd
df = pd.read_csv('/kaggle/input/reliance-nse-stock-data/RILO - Copy.csv')
df.sample(3)
y = df.iloc[:, 9]
df.shape
df.isnull().sum() | code |
121151894/cell_18 | [
"text_html_output_1.png"
] | import pandas as pd
df = pd.read_csv('/kaggle/input/reliance-nse-stock-data/RILO - Copy.csv')
df2 = pd.read_csv('/kaggle/input/reliance-nse-stock-data/RILO - Copy.csv', usecols=['Open Price', 'Low Price', 'Close Price', 'Deliverable Qty', '% Dly Qt to Traded Qty', 'Average Price'])
df2.head() | code |
121151894/cell_28 | [
"text_plain_output_1.png"
] | from sklearn.linear_model import LinearRegression,LogisticRegression
from sklearn.metrics import accuracy_score,r2_score
from sklearn.preprocessing import StandardScaler,MinMaxScaler
ss = StandardScaler()
x_train_ss = ss.fit_transform(x_train)
x_test_ss = ss.fit_transform(x_test)
lr2 = LinearRegression()
lr2.fit(x_train_ss, y_train)
y_pred1 = lr2.predict(x_test_ss)
r2_score(y_pred1, y_test) | code |
121151894/cell_8 | [
"text_plain_output_1.png"
] | import pandas as pd
df = pd.read_csv('/kaggle/input/reliance-nse-stock-data/RILO - Copy.csv')
df.sample(3)
y = df.iloc[:, 9]
df.shape
df.isnull().sum()
df.info() | code |
121151894/cell_15 | [
"text_plain_output_1.png"
] | from sklearn.linear_model import LinearRegression,LogisticRegression
from sklearn.model_selection import train_test_split,cross_val_score
lr = LinearRegression()
lr.fit(x_train, y_train)
lr.predict(x_test)
y_pred = lr.predict(x_test)
cross_val_score(lr, x_train, y_train, scoring='r2').mean() | code |
121151894/cell_3 | [
"text_plain_output_1.png"
] | import pandas as pd
df = pd.read_csv('/kaggle/input/reliance-nse-stock-data/RILO - Copy.csv')
df.sample(3) | code |
121151894/cell_14 | [
"text_plain_output_1.png"
] | from sklearn.linear_model import LinearRegression,LogisticRegression
from sklearn.metrics import accuracy_score,r2_score
lr = LinearRegression()
lr.fit(x_train, y_train)
lr.predict(x_test)
y_pred = lr.predict(x_test)
r2_score(y_pred, y_test) | code |
121151894/cell_10 | [
"text_html_output_1.png"
] | x_train.head() | code |
121151894/cell_12 | [
"text_plain_output_1.png"
] | from sklearn.linear_model import LinearRegression,LogisticRegression
lr = LinearRegression()
lr.fit(x_train, y_train)
lr.predict(x_test) | code |
128024365/cell_9 | [
"text_plain_output_1.png",
"image_output_1.png"
] | import pandas as pd
stimulus=pd.read_excel("/kaggle/input/young-adults-affective-data-ecg-and-gsr-signals/ECG_GSR_Emotions/Stimulus_Description.xlsx")
stimulus["Target Emotion"]=stimulus["Target Emotion"].str.title()
stimulus.info()
stimulus.head()
gsr_data = pd.read_excel('/kaggle/input/young-adults-affective-data-ecg-and-gsr-signals/ECG_GSR_Emotions/Self-Annotation Labels/Self-annotation Multimodal_Use.xlsx')
gsr_data['filename'] = 'GSRdata_s' + gsr_data['Session ID'].astype(str) + 'p' + gsr_data['Participant Id'].astype(str) + 'v' + gsr_data['Video ID'].astype(str) + '.dat'
PATH_2 = '/kaggle/input/young-adults-affective-data-ecg-and-gsr-signals/ECG_GSR_Emotions/Raw Data/Multimodal/GSR/'
gsr_data['filename'] = gsr_data['filename'].str.replace('GSRdata_s2p9v3.dat', 'GSRdata_S2p9v3.dat', regex=False)
gsr_data['GSR_list'] = gsr_data['filename'].apply(lambda x: list(pd.read_table(PATH_2 + x).iloc[:, 0]))
gsr_data = gsr_data.merge(stimulus.iloc[:, 0:3], on=['Session ID', 'Video ID'])
gsr_data.to_csv('gsr_data.csv', index=False)
gsr_data = pd.read_csv('/kaggle/working/gsr_data.csv')
gsr_data.info()
gsr_data.head() | code |
128024365/cell_23 | [
"image_output_1.png"
] | from sklearn.ensemble import RandomForestClassifier
from sklearn.metrics import classification_report,confusion_matrix,accuracy_score,roc_auc_score,roc_curve,ConfusionMatrixDisplay,RocCurveDisplay
from sklearn.model_selection import train_test_split
import ast
import matplotlib.pyplot as plt
import pandas as pd
stimulus=pd.read_excel("/kaggle/input/young-adults-affective-data-ecg-and-gsr-signals/ECG_GSR_Emotions/Stimulus_Description.xlsx")
stimulus["Target Emotion"]=stimulus["Target Emotion"].str.title()
stimulus.info()
stimulus.head()
gsr_data = pd.read_excel('/kaggle/input/young-adults-affective-data-ecg-and-gsr-signals/ECG_GSR_Emotions/Self-Annotation Labels/Self-annotation Multimodal_Use.xlsx')
gsr_data['filename'] = 'GSRdata_s' + gsr_data['Session ID'].astype(str) + 'p' + gsr_data['Participant Id'].astype(str) + 'v' + gsr_data['Video ID'].astype(str) + '.dat'
PATH_2 = '/kaggle/input/young-adults-affective-data-ecg-and-gsr-signals/ECG_GSR_Emotions/Raw Data/Multimodal/GSR/'
gsr_data['filename'] = gsr_data['filename'].str.replace('GSRdata_s2p9v3.dat', 'GSRdata_S2p9v3.dat', regex=False)
gsr_data['GSR_list'] = gsr_data['filename'].apply(lambda x: list(pd.read_table(PATH_2 + x).iloc[:, 0]))
gsr_data = gsr_data.merge(stimulus.iloc[:, 0:3], on=['Session ID', 'Video ID'])
gsr_data.to_csv('gsr_data.csv', index=False)
gsr_data = pd.read_csv('/kaggle/working/gsr_data.csv')
gsr_proc = gsr_data[['Emotion', 'GSR_list']].copy()
gsr_proc['GSR_list'] = gsr_proc['GSR_list'].apply(lambda x: ast.literal_eval(x))
gsr_proc = pd.concat([gsr_proc, pd.DataFrame(gsr_proc['GSR_list'].tolist())], axis=1).drop(columns=['GSR_list'])
for i in range(gsr_proc.shape[1] - 1):
gsr_proc[i] = pd.to_numeric(gsr_proc[i])
gsr_proc.to_csv('gsr_proc.csv', index=False)
gsr_proc = pd.read_csv('/kaggle/working/gsr_proc.csv')
emos = ['Happy', 'Mixed', 'Sad', 'Surprise', 'Fear', 'Disgust', 'Neutral', 'Anger']
cm = confusion_matrix(gsr_data['Target Emotion'], gsr_data['Emotion'], labels=emos)
def plot_signals(data_arr, title=''):
plt.clf()
for index, row in data_arr.iterrows():
y = row
plt.tight_layout()
plt
def model_eval(model, proc_data, label_name):
train_data, test_data = train_test_split(proc_data, test_size=0.3, random_state=3)
model.fit(train_data.drop(columns=[label_name]), train_data[label_name])
y_pred = model.predict(test_data.drop([label_name], axis=1))
emos = ['Happy', 'Mixed', 'Sad', 'Surprise', 'Fear', 'Disgust', 'Neutral', 'Anger']
cm = confusion_matrix(test_data[label_name], y_pred, labels=emos)
model_eval(RandomForestClassifier(), gsr_proc, 'Emotion') | code |
128024365/cell_15 | [
"text_html_output_1.png",
"text_plain_output_1.png"
] | from sklearn.metrics import classification_report,confusion_matrix,accuracy_score,roc_auc_score,roc_curve,ConfusionMatrixDisplay,RocCurveDisplay
import matplotlib.pyplot as plt
import pandas as pd
stimulus=pd.read_excel("/kaggle/input/young-adults-affective-data-ecg-and-gsr-signals/ECG_GSR_Emotions/Stimulus_Description.xlsx")
stimulus["Target Emotion"]=stimulus["Target Emotion"].str.title()
stimulus.info()
stimulus.head()
gsr_data = pd.read_excel('/kaggle/input/young-adults-affective-data-ecg-and-gsr-signals/ECG_GSR_Emotions/Self-Annotation Labels/Self-annotation Multimodal_Use.xlsx')
gsr_data['filename'] = 'GSRdata_s' + gsr_data['Session ID'].astype(str) + 'p' + gsr_data['Participant Id'].astype(str) + 'v' + gsr_data['Video ID'].astype(str) + '.dat'
PATH_2 = '/kaggle/input/young-adults-affective-data-ecg-and-gsr-signals/ECG_GSR_Emotions/Raw Data/Multimodal/GSR/'
gsr_data['filename'] = gsr_data['filename'].str.replace('GSRdata_s2p9v3.dat', 'GSRdata_S2p9v3.dat', regex=False)
gsr_data['GSR_list'] = gsr_data['filename'].apply(lambda x: list(pd.read_table(PATH_2 + x).iloc[:, 0]))
gsr_data = gsr_data.merge(stimulus.iloc[:, 0:3], on=['Session ID', 'Video ID'])
gsr_data.to_csv('gsr_data.csv', index=False)
gsr_data = pd.read_csv('/kaggle/working/gsr_data.csv')
emos = ['Happy', 'Mixed', 'Sad', 'Surprise', 'Fear', 'Disgust', 'Neutral', 'Anger']
cm = confusion_matrix(gsr_data['Target Emotion'], gsr_data['Emotion'], labels=emos)
ConfusionMatrixDisplay(confusion_matrix=cm, display_labels=emos).plot(cmap='YlGnBu')
plt.ylabel('Intended Video Emotion', fontsize=14)
plt.xlabel("Subject's Emotion after Watching", fontsize=14)
plt.show() | code |
128024365/cell_17 | [
"text_html_output_1.png",
"text_plain_output_1.png"
] | from sklearn.metrics import classification_report,confusion_matrix,accuracy_score,roc_auc_score,roc_curve,ConfusionMatrixDisplay,RocCurveDisplay
import ast
import matplotlib.pyplot as plt
import pandas as pd
stimulus=pd.read_excel("/kaggle/input/young-adults-affective-data-ecg-and-gsr-signals/ECG_GSR_Emotions/Stimulus_Description.xlsx")
stimulus["Target Emotion"]=stimulus["Target Emotion"].str.title()
stimulus.info()
stimulus.head()
gsr_data = pd.read_excel('/kaggle/input/young-adults-affective-data-ecg-and-gsr-signals/ECG_GSR_Emotions/Self-Annotation Labels/Self-annotation Multimodal_Use.xlsx')
gsr_data['filename'] = 'GSRdata_s' + gsr_data['Session ID'].astype(str) + 'p' + gsr_data['Participant Id'].astype(str) + 'v' + gsr_data['Video ID'].astype(str) + '.dat'
PATH_2 = '/kaggle/input/young-adults-affective-data-ecg-and-gsr-signals/ECG_GSR_Emotions/Raw Data/Multimodal/GSR/'
gsr_data['filename'] = gsr_data['filename'].str.replace('GSRdata_s2p9v3.dat', 'GSRdata_S2p9v3.dat', regex=False)
gsr_data['GSR_list'] = gsr_data['filename'].apply(lambda x: list(pd.read_table(PATH_2 + x).iloc[:, 0]))
gsr_data = gsr_data.merge(stimulus.iloc[:, 0:3], on=['Session ID', 'Video ID'])
gsr_data.to_csv('gsr_data.csv', index=False)
gsr_data = pd.read_csv('/kaggle/working/gsr_data.csv')
gsr_proc = gsr_data[['Emotion', 'GSR_list']].copy()
gsr_proc['GSR_list'] = gsr_proc['GSR_list'].apply(lambda x: ast.literal_eval(x))
gsr_proc = pd.concat([gsr_proc, pd.DataFrame(gsr_proc['GSR_list'].tolist())], axis=1).drop(columns=['GSR_list'])
for i in range(gsr_proc.shape[1] - 1):
gsr_proc[i] = pd.to_numeric(gsr_proc[i])
gsr_proc.to_csv('gsr_proc.csv', index=False)
gsr_proc = pd.read_csv('/kaggle/working/gsr_proc.csv')
emos = ['Happy', 'Mixed', 'Sad', 'Surprise', 'Fear', 'Disgust', 'Neutral', 'Anger']
cm = confusion_matrix(gsr_data['Target Emotion'], gsr_data['Emotion'], labels=emos)
def plot_signals(data_arr, title=''):
plt.clf()
plt.figure(figsize=(12, 4))
for index, row in data_arr.iterrows():
y = row
plt.plot(y)
plt.tight_layout()
plt.title(title)
plt
plt.show()
for i in emos:
plot_signals(gsr_proc[gsr_proc['Emotion'] == i].iloc[:, 1:4998:500], 'GSR Signal - ' + i) | code |
128024365/cell_12 | [
"text_html_output_1.png",
"text_plain_output_1.png"
] | import ast
import pandas as pd
stimulus=pd.read_excel("/kaggle/input/young-adults-affective-data-ecg-and-gsr-signals/ECG_GSR_Emotions/Stimulus_Description.xlsx")
stimulus["Target Emotion"]=stimulus["Target Emotion"].str.title()
stimulus.info()
stimulus.head()
gsr_data = pd.read_excel('/kaggle/input/young-adults-affective-data-ecg-and-gsr-signals/ECG_GSR_Emotions/Self-Annotation Labels/Self-annotation Multimodal_Use.xlsx')
gsr_data['filename'] = 'GSRdata_s' + gsr_data['Session ID'].astype(str) + 'p' + gsr_data['Participant Id'].astype(str) + 'v' + gsr_data['Video ID'].astype(str) + '.dat'
PATH_2 = '/kaggle/input/young-adults-affective-data-ecg-and-gsr-signals/ECG_GSR_Emotions/Raw Data/Multimodal/GSR/'
gsr_data['filename'] = gsr_data['filename'].str.replace('GSRdata_s2p9v3.dat', 'GSRdata_S2p9v3.dat', regex=False)
gsr_data['GSR_list'] = gsr_data['filename'].apply(lambda x: list(pd.read_table(PATH_2 + x).iloc[:, 0]))
gsr_data = gsr_data.merge(stimulus.iloc[:, 0:3], on=['Session ID', 'Video ID'])
gsr_data.to_csv('gsr_data.csv', index=False)
gsr_data = pd.read_csv('/kaggle/working/gsr_data.csv')
gsr_proc = gsr_data[['Emotion', 'GSR_list']].copy()
gsr_proc['GSR_list'] = gsr_proc['GSR_list'].apply(lambda x: ast.literal_eval(x))
gsr_proc = pd.concat([gsr_proc, pd.DataFrame(gsr_proc['GSR_list'].tolist())], axis=1).drop(columns=['GSR_list'])
for i in range(gsr_proc.shape[1] - 1):
gsr_proc[i] = pd.to_numeric(gsr_proc[i])
gsr_proc.to_csv('gsr_proc.csv', index=False)
gsr_proc = pd.read_csv('/kaggle/working/gsr_proc.csv')
gsr_proc.info()
gsr_proc.head() | code |
128024365/cell_5 | [
"text_plain_output_5.png",
"text_plain_output_4.png",
"image_output_5.png",
"text_plain_output_6.png",
"image_output_7.png",
"text_plain_output_3.png",
"image_output_4.png",
"text_plain_output_7.png",
"image_output_8.png",
"text_plain_output_8.png",
"image_output_6.png",
"text_plain_output_2.png",
"text_plain_output_1.png",
"image_output_3.png",
"image_output_2.png",
"image_output_1.png"
] | import pandas as pd
stimulus = pd.read_excel('/kaggle/input/young-adults-affective-data-ecg-and-gsr-signals/ECG_GSR_Emotions/Stimulus_Description.xlsx')
stimulus['Target Emotion'] = stimulus['Target Emotion'].str.title()
stimulus.info()
stimulus.head() | code |
88076802/cell_13 | [
"text_html_output_1.png"
] | import pandas as pd
import seaborn as sns
train_data = pd.read_csv('../input/nlp-getting-started/train.csv')
test_data = pd.read_csv('../input/nlp-getting-started/test.csv')
sns.countplot(x=val_df['target'], palette='vlag') | code |
88076802/cell_25 | [
"text_plain_output_1.png"
] | import torch
device = 'cuda' if torch.cuda.is_available() else 'cpu'
print(device) | code |
88076802/cell_4 | [
"application_vnd.jupyter.stderr_output_2.png",
"application_vnd.jupyter.stderr_output_1.png"
] | import pandas as pd
import seaborn as sns
train_data = pd.read_csv('../input/nlp-getting-started/train.csv')
test_data = pd.read_csv('../input/nlp-getting-started/test.csv')
sns.countplot(x=train_data['target'], palette='vlag') | code |
88076802/cell_23 | [
"text_plain_output_1.png"
] | from datasets import load_dataset
from transformers import DataCollatorWithPadding, BertForSequenceClassification, AutoModelForSequenceClassification, AutoTokenizer, Trainer, TrainingArguments
train_dataset = load_dataset('csv', data_files='train_data_clean.csv')
val_dataset = load_dataset('csv', data_files='val_data_clean.csv')
test_dataset = load_dataset('csv', data_files='test_data_clean.csv')
train_dataset = train_dataset.rename_column('text', 'sentence1')
train_dataset = train_dataset.rename_column('target', 'label')
val_dataset = val_dataset.rename_column('text', 'sentence1')
val_dataset = val_dataset.rename_column('target', 'label')
test_dataset = test_dataset.rename_column('text', 'sentence1')
checkpoint = '../input/transformers/roberta-base'
tokenizer = AutoTokenizer.from_pretrained(checkpoint)
data_collator = DataCollatorWithPadding(tokenizer=tokenizer)
def tokenize_function(dset):
return tokenizer(dset['sentence1'])
tokenized_train_dataset = train_dataset.map(tokenize_function, batched=True)
tokenized_val_dataset = val_dataset.map(tokenize_function, batched=True)
tokenized_test_dataset = test_dataset.map(tokenize_function, batched=True)
tokenized_train_dataset | code |
88076802/cell_30 | [
"text_plain_output_1.png"
] | from datasets import load_dataset
from sklearn.metrics import f1_score, accuracy_score
from transformers import DataCollatorWithPadding, BertForSequenceClassification, AutoModelForSequenceClassification, AutoTokenizer, Trainer, TrainingArguments
import numpy as np
import torch
train_dataset = load_dataset('csv', data_files='train_data_clean.csv')
val_dataset = load_dataset('csv', data_files='val_data_clean.csv')
test_dataset = load_dataset('csv', data_files='test_data_clean.csv')
train_dataset = train_dataset.rename_column('text', 'sentence1')
train_dataset = train_dataset.rename_column('target', 'label')
val_dataset = val_dataset.rename_column('text', 'sentence1')
val_dataset = val_dataset.rename_column('target', 'label')
test_dataset = test_dataset.rename_column('text', 'sentence1')
checkpoint = '../input/transformers/roberta-base'
tokenizer = AutoTokenizer.from_pretrained(checkpoint)
data_collator = DataCollatorWithPadding(tokenizer=tokenizer)
def tokenize_function(dset):
return tokenizer(dset['sentence1'])
tokenized_train_dataset = train_dataset.map(tokenize_function, batched=True)
tokenized_val_dataset = val_dataset.map(tokenize_function, batched=True)
tokenized_test_dataset = test_dataset.map(tokenize_function, batched=True)
device = 'cuda' if torch.cuda.is_available() else 'cpu'
model = AutoModelForSequenceClassification.from_pretrained(checkpoint, num_labels=2).to(device)
def compute_metrics(p):
pred, labels = p
pred = np.argmax(pred, axis=1)
f1 = f1_score(y_true=labels, y_pred=pred)
return {'f1_score': f1}
training_args = TrainingArguments(output_dir='trainer_dir', report_to='all', num_train_epochs=3)
trainer = Trainer(model, training_args, train_dataset=tokenized_train_dataset['train'], eval_dataset=tokenized_val_dataset['train'], data_collator=data_collator, tokenizer=tokenizer, compute_metrics=compute_metrics)
trainer.train() | code |
88076802/cell_29 | [
"application_vnd.jupyter.stderr_output_1.png"
] | from datasets import load_dataset
from sklearn.metrics import f1_score, accuracy_score
from transformers import DataCollatorWithPadding, BertForSequenceClassification, AutoModelForSequenceClassification, AutoTokenizer, Trainer, TrainingArguments
import numpy as np
import torch
train_dataset = load_dataset('csv', data_files='train_data_clean.csv')
val_dataset = load_dataset('csv', data_files='val_data_clean.csv')
test_dataset = load_dataset('csv', data_files='test_data_clean.csv')
train_dataset = train_dataset.rename_column('text', 'sentence1')
train_dataset = train_dataset.rename_column('target', 'label')
val_dataset = val_dataset.rename_column('text', 'sentence1')
val_dataset = val_dataset.rename_column('target', 'label')
test_dataset = test_dataset.rename_column('text', 'sentence1')
checkpoint = '../input/transformers/roberta-base'
tokenizer = AutoTokenizer.from_pretrained(checkpoint)
data_collator = DataCollatorWithPadding(tokenizer=tokenizer)
def tokenize_function(dset):
return tokenizer(dset['sentence1'])
tokenized_train_dataset = train_dataset.map(tokenize_function, batched=True)
tokenized_val_dataset = val_dataset.map(tokenize_function, batched=True)
tokenized_test_dataset = test_dataset.map(tokenize_function, batched=True)
device = 'cuda' if torch.cuda.is_available() else 'cpu'
model = AutoModelForSequenceClassification.from_pretrained(checkpoint, num_labels=2).to(device)
def compute_metrics(p):
pred, labels = p
pred = np.argmax(pred, axis=1)
f1 = f1_score(y_true=labels, y_pred=pred)
return {'f1_score': f1}
training_args = TrainingArguments(output_dir='trainer_dir', report_to='all', num_train_epochs=3)
trainer = Trainer(model, training_args, train_dataset=tokenized_train_dataset['train'], eval_dataset=tokenized_val_dataset['train'], data_collator=data_collator, tokenizer=tokenizer, compute_metrics=compute_metrics) | code |
88076802/cell_26 | [
"text_plain_output_1.png"
] | from transformers import DataCollatorWithPadding, BertForSequenceClassification, AutoModelForSequenceClassification, AutoTokenizer, Trainer, TrainingArguments
import torch
checkpoint = '../input/transformers/roberta-base'
device = 'cuda' if torch.cuda.is_available() else 'cpu'
model = AutoModelForSequenceClassification.from_pretrained(checkpoint, num_labels=2).to(device) | code |
88076802/cell_18 | [
"text_plain_output_5.png",
"text_plain_output_4.png",
"text_plain_output_6.png",
"text_plain_output_3.png",
"text_plain_output_2.png",
"text_plain_output_1.png"
] | from datasets import load_dataset
train_dataset = load_dataset('csv', data_files='train_data_clean.csv')
val_dataset = load_dataset('csv', data_files='val_data_clean.csv')
test_dataset = load_dataset('csv', data_files='test_data_clean.csv')
train_dataset = train_dataset.rename_column('text', 'sentence1')
train_dataset = train_dataset.rename_column('target', 'label')
val_dataset = val_dataset.rename_column('text', 'sentence1')
val_dataset = val_dataset.rename_column('target', 'label')
test_dataset = test_dataset.rename_column('text', 'sentence1')
val_dataset | code |
88076802/cell_15 | [
"text_plain_output_1.png",
"image_output_1.png"
] | from datasets import load_dataset
train_dataset = load_dataset('csv', data_files='train_data_clean.csv')
val_dataset = load_dataset('csv', data_files='val_data_clean.csv')
test_dataset = load_dataset('csv', data_files='test_data_clean.csv') | code |
88076802/cell_16 | [
"text_plain_output_1.png",
"image_output_1.png"
] | from datasets import load_dataset
train_dataset = load_dataset('csv', data_files='train_data_clean.csv')
val_dataset = load_dataset('csv', data_files='val_data_clean.csv')
test_dataset = load_dataset('csv', data_files='test_data_clean.csv')
val_dataset | code |
88076802/cell_3 | [
"text_plain_output_1.png"
] | import pandas as pd
train_data = pd.read_csv('../input/nlp-getting-started/train.csv')
test_data = pd.read_csv('../input/nlp-getting-started/test.csv')
train_data.tail() | code |
88076802/cell_22 | [
"text_plain_output_1.png"
] | from datasets import load_dataset
from transformers import DataCollatorWithPadding, BertForSequenceClassification, AutoModelForSequenceClassification, AutoTokenizer, Trainer, TrainingArguments
train_dataset = load_dataset('csv', data_files='train_data_clean.csv')
val_dataset = load_dataset('csv', data_files='val_data_clean.csv')
test_dataset = load_dataset('csv', data_files='test_data_clean.csv')
train_dataset = train_dataset.rename_column('text', 'sentence1')
train_dataset = train_dataset.rename_column('target', 'label')
val_dataset = val_dataset.rename_column('text', 'sentence1')
val_dataset = val_dataset.rename_column('target', 'label')
test_dataset = test_dataset.rename_column('text', 'sentence1')
checkpoint = '../input/transformers/roberta-base'
tokenizer = AutoTokenizer.from_pretrained(checkpoint)
data_collator = DataCollatorWithPadding(tokenizer=tokenizer)
def tokenize_function(dset):
return tokenizer(dset['sentence1'])
tokenized_train_dataset = train_dataset.map(tokenize_function, batched=True)
tokenized_val_dataset = val_dataset.map(tokenize_function, batched=True)
tokenized_test_dataset = test_dataset.map(tokenize_function, batched=True) | code |
88076802/cell_10 | [
"text_html_output_1.png"
] | import pandas as pd
train_data = pd.read_csv('../input/nlp-getting-started/train.csv')
test_data = pd.read_csv('../input/nlp-getting-started/test.csv')
train_data = train_data.drop(['keyword', 'location'], axis=1)
test_data = test_data.drop(['keyword', 'location'], axis=1)
train_data.tail() | code |
88076802/cell_27 | [
"text_plain_output_1.png"
] | from transformers import DataCollatorWithPadding, BertForSequenceClassification, AutoModelForSequenceClassification, AutoTokenizer, Trainer, TrainingArguments
import torch
checkpoint = '../input/transformers/roberta-base'
device = 'cuda' if torch.cuda.is_available() else 'cpu'
model = AutoModelForSequenceClassification.from_pretrained(checkpoint, num_labels=2).to(device)
model | code |
88076802/cell_12 | [
"text_plain_output_1.png",
"image_output_1.png"
] | import pandas as pd
import seaborn as sns
train_data = pd.read_csv('../input/nlp-getting-started/train.csv')
test_data = pd.read_csv('../input/nlp-getting-started/test.csv')
sns.countplot(x=train_df['target'], palette='vlag') | code |
89141106/cell_63 | [
"image_output_1.png"
] | from matplotlib import style
from matplotlib.gridspec import GridSpec
import matplotlib.pyplot as plt
import numpy as np
import pandas as pd
import seaborn as sns
arquivo = '../input/heart-failure-prediction/heart.csv'
dados = pd.read_csv(arquivo)
dados
trocar_nomes = {'Age': 'Idade', 'Sex': 'Sexo', 'ChestPainType': 'Tipo de dor', 'RestingBP': 'Pressão', 'Cholesterol': 'Colesterol', 'FastingBS': 'Glicemia', 'RestingECG': 'Eletro', 'MaxHR': 'BPM', 'ExerciseAngina': 'Dor por exec.', 'ST_Slope': 'Incl. ST', 'HeartDisease': 'DCV'}
dados = dados.rename(columns=trocar_nomes)
dados
dados.describe().T
dados = dados[dados['Colesterol'] != 0]
dados = dados[dados['Pressão'] != 0]
dados.describe().T
sns.set_theme()
style.use('fivethirtyeight')
cores = ['lightcoral', 'deepskyblue', 'orchid', 'tomato', 'teal', 'darkcyan', 'limegreen', 'darkorange']
# Definir a função do gráfico de pizza
def grafico_pizza(data_frame, coluna, cores, explode, titulo, fonte):
# Fazer contagem dos valores da coluna selecionada
df = data_frame[coluna].value_counts()
# Determinar o tamannho da plotagem
plt.figure(figsize=(15, 10))
# Criar o gráfico de pizza
_, _, pacotes = plt.pie(df,
colors=cores,
labels=df.index,
explode=explode,
shadow=True,
startangle=90,
autopct='%1.1f%%',
textprops={'fontsize': fonte,
'color': 'black',
'weight': 'bold',
'family': 'serif'})
# Plotar o gráfico de pizza
plt.setp(pacotes, color='white')
# Colocar o título do gráfico
plt.title(titulo, size=45)
# Desenhar o círculo interno
circulo_centro = plt.Circle((0, 0), 0.40, fc='white')
fig = plt.gcf()
fig.gca().add_artist(circulo_centro)
# Definir o gráfico da função de distribuição
def grafico_distribuicao(data_frame, coluna, titulo):
# Armazenar os dados da coluna
dados = data_frame[coluna]
# Determinar a figura e seu tamanho
fig = plt.figure(figsize=(17, 7))
# Criar a grade em que os gráficos serão plotados
grade = GridSpec(nrows=2, ncols=1, figure=fig)
# Escolher uma das cores para o gráfico
cor = np.random.choice(cores, 1)[0]
# Motrar o valor de assimetria dos dados
print(f'Assimetria de {titulo}: {np.round(dados.skew(), 3)}')
# Plotar o histograma
ax0 = fig.add_subplot(grade[0, :])
ax0.set_title(f'Histograma e BoxPlot de {titulo}', y=1.05)
sns.histplot(data=dados, ax=ax0, color=cor)
# Plotar o BoxPlot
ax1 = fig.add_subplot(grade[1, :])
plt.axis('off')
sns.boxplot(x=dados, ax=ax1, color=cor)
# Definir o gráfico de influência
def grafico_influencia(data_frame, coluna, bins, labels, com_bins=True):
# Armazenar os dados da coluna
influencia = data_frame.loc[:, [coluna, 'DCV']]
# Se os dados de "coluna" não forem como classe, então terá intervalos ("bins" e "labels")
if com_bins:
influencia[coluna] = pd.cut(influencia[coluna],
bins=bins,
labels=labels)
# Escolher uma das cores para o gráfico
cor = np.random.choice(cores, 1)[0]
# Determinar o tamanho da figura
plt.figure(figsize=(15, 5))
# Criar o gráfico
grafico = sns.pointplot(x=coluna, y='DCV', dodge=0.1, capsize=.1, data=influencia, color=cor)
# Colocar o título do gráfico
grafico.set_title(f'{coluna} influência', fontsize=25)
grafico_influencia(dados, 'Glicemia', None, None, False) | code |
89141106/cell_21 | [
"image_output_1.png"
] | import pandas as pd
arquivo = '../input/heart-failure-prediction/heart.csv'
dados = pd.read_csv(arquivo)
dados
trocar_nomes = {'Age': 'Idade', 'Sex': 'Sexo', 'ChestPainType': 'Tipo de dor', 'RestingBP': 'Pressão', 'Cholesterol': 'Colesterol', 'FastingBS': 'Glicemia', 'RestingECG': 'Eletro', 'MaxHR': 'BPM', 'ExerciseAngina': 'Dor por exec.', 'ST_Slope': 'Incl. ST', 'HeartDisease': 'DCV'}
dados = dados.rename(columns=trocar_nomes)
dados
dados.describe().T | code |
89141106/cell_25 | [
"image_output_1.png"
] | import pandas as pd
arquivo = '../input/heart-failure-prediction/heart.csv'
dados = pd.read_csv(arquivo)
dados
trocar_nomes = {'Age': 'Idade', 'Sex': 'Sexo', 'ChestPainType': 'Tipo de dor', 'RestingBP': 'Pressão', 'Cholesterol': 'Colesterol', 'FastingBS': 'Glicemia', 'RestingECG': 'Eletro', 'MaxHR': 'BPM', 'ExerciseAngina': 'Dor por exec.', 'ST_Slope': 'Incl. ST', 'HeartDisease': 'DCV'}
dados = dados.rename(columns=trocar_nomes)
dados
dados.describe().T
dados = dados[dados['Colesterol'] != 0]
dados = dados[dados['Pressão'] != 0]
dados.describe().T | code |
89141106/cell_83 | [
"text_plain_output_1.png",
"image_output_1.png"
] | from matplotlib import style
from matplotlib.gridspec import GridSpec
import matplotlib.pyplot as plt
import numpy as np
import pandas as pd
import seaborn as sns
arquivo = '../input/heart-failure-prediction/heart.csv'
dados = pd.read_csv(arquivo)
dados
trocar_nomes = {'Age': 'Idade', 'Sex': 'Sexo', 'ChestPainType': 'Tipo de dor', 'RestingBP': 'Pressão', 'Cholesterol': 'Colesterol', 'FastingBS': 'Glicemia', 'RestingECG': 'Eletro', 'MaxHR': 'BPM', 'ExerciseAngina': 'Dor por exec.', 'ST_Slope': 'Incl. ST', 'HeartDisease': 'DCV'}
dados = dados.rename(columns=trocar_nomes)
dados
dados.describe().T
dados = dados[dados['Colesterol'] != 0]
dados = dados[dados['Pressão'] != 0]
dados.describe().T
sns.set_theme()
style.use('fivethirtyeight')
cores = ['lightcoral', 'deepskyblue', 'orchid', 'tomato', 'teal', 'darkcyan', 'limegreen', 'darkorange']
# Definir a função do gráfico de pizza
def grafico_pizza(data_frame, coluna, cores, explode, titulo, fonte):
# Fazer contagem dos valores da coluna selecionada
df = data_frame[coluna].value_counts()
# Determinar o tamannho da plotagem
plt.figure(figsize=(15, 10))
# Criar o gráfico de pizza
_, _, pacotes = plt.pie(df,
colors=cores,
labels=df.index,
explode=explode,
shadow=True,
startangle=90,
autopct='%1.1f%%',
textprops={'fontsize': fonte,
'color': 'black',
'weight': 'bold',
'family': 'serif'})
# Plotar o gráfico de pizza
plt.setp(pacotes, color='white')
# Colocar o título do gráfico
plt.title(titulo, size=45)
# Desenhar o círculo interno
circulo_centro = plt.Circle((0, 0), 0.40, fc='white')
fig = plt.gcf()
fig.gca().add_artist(circulo_centro)
# Definir o gráfico da função de distribuição
def grafico_distribuicao(data_frame, coluna, titulo):
# Armazenar os dados da coluna
dados = data_frame[coluna]
# Determinar a figura e seu tamanho
fig = plt.figure(figsize=(17, 7))
# Criar a grade em que os gráficos serão plotados
grade = GridSpec(nrows=2, ncols=1, figure=fig)
# Escolher uma das cores para o gráfico
cor = np.random.choice(cores, 1)[0]
# Motrar o valor de assimetria dos dados
print(f'Assimetria de {titulo}: {np.round(dados.skew(), 3)}')
# Plotar o histograma
ax0 = fig.add_subplot(grade[0, :])
ax0.set_title(f'Histograma e BoxPlot de {titulo}', y=1.05)
sns.histplot(data=dados, ax=ax0, color=cor)
# Plotar o BoxPlot
ax1 = fig.add_subplot(grade[1, :])
plt.axis('off')
sns.boxplot(x=dados, ax=ax1, color=cor)
# Definir o gráfico de influência
def grafico_influencia(data_frame, coluna, bins, labels, com_bins=True):
# Armazenar os dados da coluna
influencia = data_frame.loc[:, [coluna, 'DCV']]
# Se os dados de "coluna" não forem como classe, então terá intervalos ("bins" e "labels")
if com_bins:
influencia[coluna] = pd.cut(influencia[coluna],
bins=bins,
labels=labels)
# Escolher uma das cores para o gráfico
cor = np.random.choice(cores, 1)[0]
# Determinar o tamanho da figura
plt.figure(figsize=(15, 5))
# Criar o gráfico
grafico = sns.pointplot(x=coluna, y='DCV', dodge=0.1, capsize=.1, data=influencia, color=cor)
# Colocar o título do gráfico
grafico.set_title(f'{coluna} influência', fontsize=25)
grafico_influencia(dados, 'Incl. ST', None, None, False) | code |
89141106/cell_117 | [
"text_html_output_1.png"
] | from matplotlib import style
from matplotlib.gridspec import GridSpec
from sklearn import preprocessing
from sklearn.ensemble import GradientBoostingClassifier, RandomForestClassifier
from sklearn.metrics import confusion_matrix, accuracy_score, precision_score, recall_score, f1_score
from sklearn.neighbors import KNeighborsClassifier
import matplotlib.pyplot as plt
import numpy as np
import pandas as pd
import seaborn as sns
arquivo = '../input/heart-failure-prediction/heart.csv'
dados = pd.read_csv(arquivo)
dados
trocar_nomes = {'Age': 'Idade', 'Sex': 'Sexo', 'ChestPainType': 'Tipo de dor', 'RestingBP': 'Pressão', 'Cholesterol': 'Colesterol', 'FastingBS': 'Glicemia', 'RestingECG': 'Eletro', 'MaxHR': 'BPM', 'ExerciseAngina': 'Dor por exec.', 'ST_Slope': 'Incl. ST', 'HeartDisease': 'DCV'}
dados = dados.rename(columns=trocar_nomes)
dados
dados.describe().T
dados = dados[dados['Colesterol'] != 0]
dados = dados[dados['Pressão'] != 0]
dados.describe().T
colunas = ['Sexo', 'Tipo de dor', 'Glicemia', 'Eletro', 'Dor por exec.', 'Incl. ST', 'DCV']
sns.set_theme()
style.use('fivethirtyeight')
cores = ['lightcoral', 'deepskyblue', 'orchid', 'tomato', 'teal', 'darkcyan', 'limegreen', 'darkorange']
# Definir a função do gráfico de pizza
def grafico_pizza(data_frame, coluna, cores, explode, titulo, fonte):
# Fazer contagem dos valores da coluna selecionada
df = data_frame[coluna].value_counts()
# Determinar o tamannho da plotagem
plt.figure(figsize=(15, 10))
# Criar o gráfico de pizza
_, _, pacotes = plt.pie(df,
colors=cores,
labels=df.index,
explode=explode,
shadow=True,
startangle=90,
autopct='%1.1f%%',
textprops={'fontsize': fonte,
'color': 'black',
'weight': 'bold',
'family': 'serif'})
# Plotar o gráfico de pizza
plt.setp(pacotes, color='white')
# Colocar o título do gráfico
plt.title(titulo, size=45)
# Desenhar o círculo interno
circulo_centro = plt.Circle((0, 0), 0.40, fc='white')
fig = plt.gcf()
fig.gca().add_artist(circulo_centro)
# Definir o gráfico da função de distribuição
def grafico_distribuicao(data_frame, coluna, titulo):
# Armazenar os dados da coluna
dados = data_frame[coluna]
# Determinar a figura e seu tamanho
fig = plt.figure(figsize=(17, 7))
# Criar a grade em que os gráficos serão plotados
grade = GridSpec(nrows=2, ncols=1, figure=fig)
# Escolher uma das cores para o gráfico
cor = np.random.choice(cores, 1)[0]
# Motrar o valor de assimetria dos dados
print(f'Assimetria de {titulo}: {np.round(dados.skew(), 3)}')
# Plotar o histograma
ax0 = fig.add_subplot(grade[0, :])
ax0.set_title(f'Histograma e BoxPlot de {titulo}', y=1.05)
sns.histplot(data=dados, ax=ax0, color=cor)
# Plotar o BoxPlot
ax1 = fig.add_subplot(grade[1, :])
plt.axis('off')
sns.boxplot(x=dados, ax=ax1, color=cor)
# Definir o gráfico de influência
def grafico_influencia(data_frame, coluna, bins, labels, com_bins=True):
# Armazenar os dados da coluna
influencia = data_frame.loc[:, [coluna, 'DCV']]
# Se os dados de "coluna" não forem como classe, então terá intervalos ("bins" e "labels")
if com_bins:
influencia[coluna] = pd.cut(influencia[coluna],
bins=bins,
labels=labels)
# Escolher uma das cores para o gráfico
cor = np.random.choice(cores, 1)[0]
# Determinar o tamanho da figura
plt.figure(figsize=(15, 5))
# Criar o gráfico
grafico = sns.pointplot(x=coluna, y='DCV', dodge=0.1, capsize=.1, data=influencia, color=cor)
# Colocar o título do gráfico
grafico.set_title(f'{coluna} influência', fontsize=25)
colunas = [coluna for coluna in dados.columns if dados[coluna].dtype == 'object']
codificador = preprocessing.LabelEncoder()
for coluna in colunas:
dados[coluna] = codificador.fit_transform(dados[coluna])
mascara = np.triu(dados.corr())
rfc = RandomForestClassifier(n_jobs=-1, n_estimators=500, max_depth=70, max_features=2, random_state=0)
knn = KNeighborsClassifier(n_neighbors=5, algorithm='kd_tree', weights='uniform', n_jobs=-1)
gbc = GradientBoostingClassifier(learning_rate=0.01, loss='exponential', max_depth=70, max_features=2, n_estimators=500, random_state=0)
knn.fit(X_treino, y_treino)
# Para avaliar as métricas e a matriz de confusão, vamos criar uma função para facilitar a nossa vida
def metricas(X_teste, y_teste, classificador, nome_clf):
# Prever o resultado
y_pred = classificador.predict(X_teste)
# Métrica de acurácia
acuracia = accuracy_score(y_teste, y_pred)
# Métrica de precisão
precisao = precision_score(y_teste, y_pred)
# Métrica de revocação
revocacao = recall_score(y_teste, y_pred)
# Métrica de pontuação F1
f1 = f1_score(y_teste, y_pred)
# Mostrar os valores das métricas
print(f'''{nome_clf}
Acurácia: {acuracia:.3f}
Precisão: {precisao:.3f}
Revocação: {revocacao:.3f}
Pontuação F1: {f1:.3f}''')
# Criar a matriz de confusão
matriz = confusion_matrix(y_teste, y_pred)
# Criar um DataFrame para aramazenar os dados de "y_pred" e "y_teste"
df_mc = pd.DataFrame(matriz, index=['DCV', 'Normal'], columns=['DCV', 'Normal'])
# Determinar o tamanho da plotagem
plt.figure(figsize=(15, 10))
# Criar a plotagem
mapa_calor = sns.heatmap(data=df_mc,
annot=True,
cmap='Blues',
fmt='.5g',
annot_kws={'size': 20})
# Colocar o título da matriz de confusão
mapa_calor.set_title(f'Matriz de Confusão {nome_clf}', fontsize=25, y=1.05)
# Adicionar a legenda nos eixos
plt.xlabel('Valores Previstos', fontsize=20)
plt.ylabel('Valores Reais', fontsize=20)
# Mostrar a matriz de confusão
plt.show()
metricas(X_teste, y_teste, knn, 'kNN') | code |
89141106/cell_79 | [
"text_plain_output_1.png"
] | from matplotlib import style
from matplotlib.gridspec import GridSpec
import matplotlib.pyplot as plt
import numpy as np
import pandas as pd
import seaborn as sns
arquivo = '../input/heart-failure-prediction/heart.csv'
dados = pd.read_csv(arquivo)
dados
trocar_nomes = {'Age': 'Idade', 'Sex': 'Sexo', 'ChestPainType': 'Tipo de dor', 'RestingBP': 'Pressão', 'Cholesterol': 'Colesterol', 'FastingBS': 'Glicemia', 'RestingECG': 'Eletro', 'MaxHR': 'BPM', 'ExerciseAngina': 'Dor por exec.', 'ST_Slope': 'Incl. ST', 'HeartDisease': 'DCV'}
dados = dados.rename(columns=trocar_nomes)
dados
dados.describe().T
dados = dados[dados['Colesterol'] != 0]
dados = dados[dados['Pressão'] != 0]
dados.describe().T
sns.set_theme()
style.use('fivethirtyeight')
cores = ['lightcoral', 'deepskyblue', 'orchid', 'tomato', 'teal', 'darkcyan', 'limegreen', 'darkorange']
# Definir a função do gráfico de pizza
def grafico_pizza(data_frame, coluna, cores, explode, titulo, fonte):
# Fazer contagem dos valores da coluna selecionada
df = data_frame[coluna].value_counts()
# Determinar o tamannho da plotagem
plt.figure(figsize=(15, 10))
# Criar o gráfico de pizza
_, _, pacotes = plt.pie(df,
colors=cores,
labels=df.index,
explode=explode,
shadow=True,
startangle=90,
autopct='%1.1f%%',
textprops={'fontsize': fonte,
'color': 'black',
'weight': 'bold',
'family': 'serif'})
# Plotar o gráfico de pizza
plt.setp(pacotes, color='white')
# Colocar o título do gráfico
plt.title(titulo, size=45)
# Desenhar o círculo interno
circulo_centro = plt.Circle((0, 0), 0.40, fc='white')
fig = plt.gcf()
fig.gca().add_artist(circulo_centro)
# Definir o gráfico da função de distribuição
def grafico_distribuicao(data_frame, coluna, titulo):
# Armazenar os dados da coluna
dados = data_frame[coluna]
# Determinar a figura e seu tamanho
fig = plt.figure(figsize=(17, 7))
# Criar a grade em que os gráficos serão plotados
grade = GridSpec(nrows=2, ncols=1, figure=fig)
# Escolher uma das cores para o gráfico
cor = np.random.choice(cores, 1)[0]
# Motrar o valor de assimetria dos dados
print(f'Assimetria de {titulo}: {np.round(dados.skew(), 3)}')
# Plotar o histograma
ax0 = fig.add_subplot(grade[0, :])
ax0.set_title(f'Histograma e BoxPlot de {titulo}', y=1.05)
sns.histplot(data=dados, ax=ax0, color=cor)
# Plotar o BoxPlot
ax1 = fig.add_subplot(grade[1, :])
plt.axis('off')
sns.boxplot(x=dados, ax=ax1, color=cor)
# Definir o gráfico de influência
def grafico_influencia(data_frame, coluna, bins, labels, com_bins=True):
# Armazenar os dados da coluna
influencia = data_frame.loc[:, [coluna, 'DCV']]
# Se os dados de "coluna" não forem como classe, então terá intervalos ("bins" e "labels")
if com_bins:
influencia[coluna] = pd.cut(influencia[coluna],
bins=bins,
labels=labels)
# Escolher uma das cores para o gráfico
cor = np.random.choice(cores, 1)[0]
# Determinar o tamanho da figura
plt.figure(figsize=(15, 5))
# Criar o gráfico
grafico = sns.pointplot(x=coluna, y='DCV', dodge=0.1, capsize=.1, data=influencia, color=cor)
# Colocar o título do gráfico
grafico.set_title(f'{coluna} influência', fontsize=25)
grafico_influencia(dados, 'Oldpeak', [-0.1, 0.3, 0.8, 1.3, 1.8, 2.3, 2.8, 3.3, 3.8, 4.3, 4.8, 7.0], ['-0.1-0.3', '0.3-0.8', '0.8-1.3', '1.3-1.8', '1.8-2.3', '2.3-2.8', '2.8-3.3', '3.3-3.8', '3.8-4.3', '4.3-4.8', '5.0+']) | code |
89141106/cell_90 | [
"text_plain_output_1.png",
"image_output_1.png"
] | from matplotlib import style
from matplotlib.gridspec import GridSpec
from sklearn import preprocessing
import matplotlib.pyplot as plt
import numpy as np
import pandas as pd
import seaborn as sns
arquivo = '../input/heart-failure-prediction/heart.csv'
dados = pd.read_csv(arquivo)
dados
trocar_nomes = {'Age': 'Idade', 'Sex': 'Sexo', 'ChestPainType': 'Tipo de dor', 'RestingBP': 'Pressão', 'Cholesterol': 'Colesterol', 'FastingBS': 'Glicemia', 'RestingECG': 'Eletro', 'MaxHR': 'BPM', 'ExerciseAngina': 'Dor por exec.', 'ST_Slope': 'Incl. ST', 'HeartDisease': 'DCV'}
dados = dados.rename(columns=trocar_nomes)
dados
dados.describe().T
dados = dados[dados['Colesterol'] != 0]
dados = dados[dados['Pressão'] != 0]
dados.describe().T
colunas = ['Sexo', 'Tipo de dor', 'Glicemia', 'Eletro', 'Dor por exec.', 'Incl. ST', 'DCV']
sns.set_theme()
style.use('fivethirtyeight')
cores = ['lightcoral', 'deepskyblue', 'orchid', 'tomato', 'teal', 'darkcyan', 'limegreen', 'darkorange']
# Definir a função do gráfico de pizza
def grafico_pizza(data_frame, coluna, cores, explode, titulo, fonte):
# Fazer contagem dos valores da coluna selecionada
df = data_frame[coluna].value_counts()
# Determinar o tamannho da plotagem
plt.figure(figsize=(15, 10))
# Criar o gráfico de pizza
_, _, pacotes = plt.pie(df,
colors=cores,
labels=df.index,
explode=explode,
shadow=True,
startangle=90,
autopct='%1.1f%%',
textprops={'fontsize': fonte,
'color': 'black',
'weight': 'bold',
'family': 'serif'})
# Plotar o gráfico de pizza
plt.setp(pacotes, color='white')
# Colocar o título do gráfico
plt.title(titulo, size=45)
# Desenhar o círculo interno
circulo_centro = plt.Circle((0, 0), 0.40, fc='white')
fig = plt.gcf()
fig.gca().add_artist(circulo_centro)
# Definir o gráfico da função de distribuição
def grafico_distribuicao(data_frame, coluna, titulo):
# Armazenar os dados da coluna
dados = data_frame[coluna]
# Determinar a figura e seu tamanho
fig = plt.figure(figsize=(17, 7))
# Criar a grade em que os gráficos serão plotados
grade = GridSpec(nrows=2, ncols=1, figure=fig)
# Escolher uma das cores para o gráfico
cor = np.random.choice(cores, 1)[0]
# Motrar o valor de assimetria dos dados
print(f'Assimetria de {titulo}: {np.round(dados.skew(), 3)}')
# Plotar o histograma
ax0 = fig.add_subplot(grade[0, :])
ax0.set_title(f'Histograma e BoxPlot de {titulo}', y=1.05)
sns.histplot(data=dados, ax=ax0, color=cor)
# Plotar o BoxPlot
ax1 = fig.add_subplot(grade[1, :])
plt.axis('off')
sns.boxplot(x=dados, ax=ax1, color=cor)
colunas = [coluna for coluna in dados.columns if dados[coluna].dtype == 'object']
codificador = preprocessing.LabelEncoder()
for coluna in colunas:
dados[coluna] = codificador.fit_transform(dados[coluna])
dados | code |
89141106/cell_44 | [
"image_output_1.png"
] | from matplotlib import style
from matplotlib.gridspec import GridSpec
import matplotlib.pyplot as plt
import numpy as np
import pandas as pd
import seaborn as sns
arquivo = '../input/heart-failure-prediction/heart.csv'
dados = pd.read_csv(arquivo)
dados
trocar_nomes = {'Age': 'Idade', 'Sex': 'Sexo', 'ChestPainType': 'Tipo de dor', 'RestingBP': 'Pressão', 'Cholesterol': 'Colesterol', 'FastingBS': 'Glicemia', 'RestingECG': 'Eletro', 'MaxHR': 'BPM', 'ExerciseAngina': 'Dor por exec.', 'ST_Slope': 'Incl. ST', 'HeartDisease': 'DCV'}
dados = dados.rename(columns=trocar_nomes)
dados
dados.describe().T
dados = dados[dados['Colesterol'] != 0]
dados = dados[dados['Pressão'] != 0]
dados.describe().T
sns.set_theme()
style.use('fivethirtyeight')
cores = ['lightcoral', 'deepskyblue', 'orchid', 'tomato', 'teal', 'darkcyan', 'limegreen', 'darkorange']
# Definir a função do gráfico de pizza
def grafico_pizza(data_frame, coluna, cores, explode, titulo, fonte):
# Fazer contagem dos valores da coluna selecionada
df = data_frame[coluna].value_counts()
# Determinar o tamannho da plotagem
plt.figure(figsize=(15, 10))
# Criar o gráfico de pizza
_, _, pacotes = plt.pie(df,
colors=cores,
labels=df.index,
explode=explode,
shadow=True,
startangle=90,
autopct='%1.1f%%',
textprops={'fontsize': fonte,
'color': 'black',
'weight': 'bold',
'family': 'serif'})
# Plotar o gráfico de pizza
plt.setp(pacotes, color='white')
# Colocar o título do gráfico
plt.title(titulo, size=45)
# Desenhar o círculo interno
circulo_centro = plt.Circle((0, 0), 0.40, fc='white')
fig = plt.gcf()
fig.gca().add_artist(circulo_centro)
# Definir o gráfico da função de distribuição
def grafico_distribuicao(data_frame, coluna, titulo):
# Armazenar os dados da coluna
dados = data_frame[coluna]
# Determinar a figura e seu tamanho
fig = plt.figure(figsize=(17, 7))
# Criar a grade em que os gráficos serão plotados
grade = GridSpec(nrows=2, ncols=1, figure=fig)
# Escolher uma das cores para o gráfico
cor = np.random.choice(cores, 1)[0]
# Motrar o valor de assimetria dos dados
print(f'Assimetria de {titulo}: {np.round(dados.skew(), 3)}')
# Plotar o histograma
ax0 = fig.add_subplot(grade[0, :])
ax0.set_title(f'Histograma e BoxPlot de {titulo}', y=1.05)
sns.histplot(data=dados, ax=ax0, color=cor)
# Plotar o BoxPlot
ax1 = fig.add_subplot(grade[1, :])
plt.axis('off')
sns.boxplot(x=dados, ax=ax1, color=cor)
grafico_distribuicao(dados, 'Idade', 'Idade') | code |
89141106/cell_55 | [
"image_output_1.png"
] | from matplotlib import style
from matplotlib.gridspec import GridSpec
import matplotlib.pyplot as plt
import numpy as np
import pandas as pd
import seaborn as sns
arquivo = '../input/heart-failure-prediction/heart.csv'
dados = pd.read_csv(arquivo)
dados
trocar_nomes = {'Age': 'Idade', 'Sex': 'Sexo', 'ChestPainType': 'Tipo de dor', 'RestingBP': 'Pressão', 'Cholesterol': 'Colesterol', 'FastingBS': 'Glicemia', 'RestingECG': 'Eletro', 'MaxHR': 'BPM', 'ExerciseAngina': 'Dor por exec.', 'ST_Slope': 'Incl. ST', 'HeartDisease': 'DCV'}
dados = dados.rename(columns=trocar_nomes)
dados
dados.describe().T
dados = dados[dados['Colesterol'] != 0]
dados = dados[dados['Pressão'] != 0]
dados.describe().T
sns.set_theme()
style.use('fivethirtyeight')
cores = ['lightcoral', 'deepskyblue', 'orchid', 'tomato', 'teal', 'darkcyan', 'limegreen', 'darkorange']
# Definir a função do gráfico de pizza
def grafico_pizza(data_frame, coluna, cores, explode, titulo, fonte):
# Fazer contagem dos valores da coluna selecionada
df = data_frame[coluna].value_counts()
# Determinar o tamannho da plotagem
plt.figure(figsize=(15, 10))
# Criar o gráfico de pizza
_, _, pacotes = plt.pie(df,
colors=cores,
labels=df.index,
explode=explode,
shadow=True,
startangle=90,
autopct='%1.1f%%',
textprops={'fontsize': fonte,
'color': 'black',
'weight': 'bold',
'family': 'serif'})
# Plotar o gráfico de pizza
plt.setp(pacotes, color='white')
# Colocar o título do gráfico
plt.title(titulo, size=45)
# Desenhar o círculo interno
circulo_centro = plt.Circle((0, 0), 0.40, fc='white')
fig = plt.gcf()
fig.gca().add_artist(circulo_centro)
# Definir o gráfico da função de distribuição
def grafico_distribuicao(data_frame, coluna, titulo):
# Armazenar os dados da coluna
dados = data_frame[coluna]
# Determinar a figura e seu tamanho
fig = plt.figure(figsize=(17, 7))
# Criar a grade em que os gráficos serão plotados
grade = GridSpec(nrows=2, ncols=1, figure=fig)
# Escolher uma das cores para o gráfico
cor = np.random.choice(cores, 1)[0]
# Motrar o valor de assimetria dos dados
print(f'Assimetria de {titulo}: {np.round(dados.skew(), 3)}')
# Plotar o histograma
ax0 = fig.add_subplot(grade[0, :])
ax0.set_title(f'Histograma e BoxPlot de {titulo}', y=1.05)
sns.histplot(data=dados, ax=ax0, color=cor)
# Plotar o BoxPlot
ax1 = fig.add_subplot(grade[1, :])
plt.axis('off')
sns.boxplot(x=dados, ax=ax1, color=cor)
# Definir o gráfico de influência
def grafico_influencia(data_frame, coluna, bins, labels, com_bins=True):
# Armazenar os dados da coluna
influencia = data_frame.loc[:, [coluna, 'DCV']]
# Se os dados de "coluna" não forem como classe, então terá intervalos ("bins" e "labels")
if com_bins:
influencia[coluna] = pd.cut(influencia[coluna],
bins=bins,
labels=labels)
# Escolher uma das cores para o gráfico
cor = np.random.choice(cores, 1)[0]
# Determinar o tamanho da figura
plt.figure(figsize=(15, 5))
# Criar o gráfico
grafico = sns.pointplot(x=coluna, y='DCV', dodge=0.1, capsize=.1, data=influencia, color=cor)
# Colocar o título do gráfico
grafico.set_title(f'{coluna} influência', fontsize=25)
grafico_influencia(dados, 'Pressão', [75, 100, 125, 150, 175, 200], ['75-100', '100-125', '125-150', '150-175', '175-200']) | code |
89141106/cell_74 | [
"text_html_output_1.png"
] | from matplotlib import style
from matplotlib.gridspec import GridSpec
import matplotlib.pyplot as plt
import numpy as np
import pandas as pd
import seaborn as sns
arquivo = '../input/heart-failure-prediction/heart.csv'
dados = pd.read_csv(arquivo)
dados
trocar_nomes = {'Age': 'Idade', 'Sex': 'Sexo', 'ChestPainType': 'Tipo de dor', 'RestingBP': 'Pressão', 'Cholesterol': 'Colesterol', 'FastingBS': 'Glicemia', 'RestingECG': 'Eletro', 'MaxHR': 'BPM', 'ExerciseAngina': 'Dor por exec.', 'ST_Slope': 'Incl. ST', 'HeartDisease': 'DCV'}
dados = dados.rename(columns=trocar_nomes)
dados
dados.describe().T
dados = dados[dados['Colesterol'] != 0]
dados = dados[dados['Pressão'] != 0]
dados.describe().T
sns.set_theme()
style.use('fivethirtyeight')
cores = ['lightcoral', 'deepskyblue', 'orchid', 'tomato', 'teal', 'darkcyan', 'limegreen', 'darkorange']
# Definir a função do gráfico de pizza
def grafico_pizza(data_frame, coluna, cores, explode, titulo, fonte):
# Fazer contagem dos valores da coluna selecionada
df = data_frame[coluna].value_counts()
# Determinar o tamannho da plotagem
plt.figure(figsize=(15, 10))
# Criar o gráfico de pizza
_, _, pacotes = plt.pie(df,
colors=cores,
labels=df.index,
explode=explode,
shadow=True,
startangle=90,
autopct='%1.1f%%',
textprops={'fontsize': fonte,
'color': 'black',
'weight': 'bold',
'family': 'serif'})
# Plotar o gráfico de pizza
plt.setp(pacotes, color='white')
# Colocar o título do gráfico
plt.title(titulo, size=45)
# Desenhar o círculo interno
circulo_centro = plt.Circle((0, 0), 0.40, fc='white')
fig = plt.gcf()
fig.gca().add_artist(circulo_centro)
# Definir o gráfico da função de distribuição
def grafico_distribuicao(data_frame, coluna, titulo):
# Armazenar os dados da coluna
dados = data_frame[coluna]
# Determinar a figura e seu tamanho
fig = plt.figure(figsize=(17, 7))
# Criar a grade em que os gráficos serão plotados
grade = GridSpec(nrows=2, ncols=1, figure=fig)
# Escolher uma das cores para o gráfico
cor = np.random.choice(cores, 1)[0]
# Motrar o valor de assimetria dos dados
print(f'Assimetria de {titulo}: {np.round(dados.skew(), 3)}')
# Plotar o histograma
ax0 = fig.add_subplot(grade[0, :])
ax0.set_title(f'Histograma e BoxPlot de {titulo}', y=1.05)
sns.histplot(data=dados, ax=ax0, color=cor)
# Plotar o BoxPlot
ax1 = fig.add_subplot(grade[1, :])
plt.axis('off')
sns.boxplot(x=dados, ax=ax1, color=cor)
grafico_pizza(dados, 'Dor por exec.', ('#140E36', '#091AAB'), (0.05, 0.05), 'Angina por exercício físico', 25) | code |
89141106/cell_116 | [
"text_html_output_1.png"
] | from matplotlib import style
from matplotlib.gridspec import GridSpec
from sklearn import preprocessing
from sklearn.ensemble import GradientBoostingClassifier, RandomForestClassifier
from sklearn.metrics import confusion_matrix, accuracy_score, precision_score, recall_score, f1_score
from sklearn.neighbors import KNeighborsClassifier
import matplotlib.pyplot as plt
import numpy as np
import pandas as pd
import seaborn as sns
arquivo = '../input/heart-failure-prediction/heart.csv'
dados = pd.read_csv(arquivo)
dados
trocar_nomes = {'Age': 'Idade', 'Sex': 'Sexo', 'ChestPainType': 'Tipo de dor', 'RestingBP': 'Pressão', 'Cholesterol': 'Colesterol', 'FastingBS': 'Glicemia', 'RestingECG': 'Eletro', 'MaxHR': 'BPM', 'ExerciseAngina': 'Dor por exec.', 'ST_Slope': 'Incl. ST', 'HeartDisease': 'DCV'}
dados = dados.rename(columns=trocar_nomes)
dados
dados.describe().T
dados = dados[dados['Colesterol'] != 0]
dados = dados[dados['Pressão'] != 0]
dados.describe().T
colunas = ['Sexo', 'Tipo de dor', 'Glicemia', 'Eletro', 'Dor por exec.', 'Incl. ST', 'DCV']
sns.set_theme()
style.use('fivethirtyeight')
cores = ['lightcoral', 'deepskyblue', 'orchid', 'tomato', 'teal', 'darkcyan', 'limegreen', 'darkorange']
# Definir a função do gráfico de pizza
def grafico_pizza(data_frame, coluna, cores, explode, titulo, fonte):
# Fazer contagem dos valores da coluna selecionada
df = data_frame[coluna].value_counts()
# Determinar o tamannho da plotagem
plt.figure(figsize=(15, 10))
# Criar o gráfico de pizza
_, _, pacotes = plt.pie(df,
colors=cores,
labels=df.index,
explode=explode,
shadow=True,
startangle=90,
autopct='%1.1f%%',
textprops={'fontsize': fonte,
'color': 'black',
'weight': 'bold',
'family': 'serif'})
# Plotar o gráfico de pizza
plt.setp(pacotes, color='white')
# Colocar o título do gráfico
plt.title(titulo, size=45)
# Desenhar o círculo interno
circulo_centro = plt.Circle((0, 0), 0.40, fc='white')
fig = plt.gcf()
fig.gca().add_artist(circulo_centro)
# Definir o gráfico da função de distribuição
def grafico_distribuicao(data_frame, coluna, titulo):
# Armazenar os dados da coluna
dados = data_frame[coluna]
# Determinar a figura e seu tamanho
fig = plt.figure(figsize=(17, 7))
# Criar a grade em que os gráficos serão plotados
grade = GridSpec(nrows=2, ncols=1, figure=fig)
# Escolher uma das cores para o gráfico
cor = np.random.choice(cores, 1)[0]
# Motrar o valor de assimetria dos dados
print(f'Assimetria de {titulo}: {np.round(dados.skew(), 3)}')
# Plotar o histograma
ax0 = fig.add_subplot(grade[0, :])
ax0.set_title(f'Histograma e BoxPlot de {titulo}', y=1.05)
sns.histplot(data=dados, ax=ax0, color=cor)
# Plotar o BoxPlot
ax1 = fig.add_subplot(grade[1, :])
plt.axis('off')
sns.boxplot(x=dados, ax=ax1, color=cor)
# Definir o gráfico de influência
def grafico_influencia(data_frame, coluna, bins, labels, com_bins=True):
# Armazenar os dados da coluna
influencia = data_frame.loc[:, [coluna, 'DCV']]
# Se os dados de "coluna" não forem como classe, então terá intervalos ("bins" e "labels")
if com_bins:
influencia[coluna] = pd.cut(influencia[coluna],
bins=bins,
labels=labels)
# Escolher uma das cores para o gráfico
cor = np.random.choice(cores, 1)[0]
# Determinar o tamanho da figura
plt.figure(figsize=(15, 5))
# Criar o gráfico
grafico = sns.pointplot(x=coluna, y='DCV', dodge=0.1, capsize=.1, data=influencia, color=cor)
# Colocar o título do gráfico
grafico.set_title(f'{coluna} influência', fontsize=25)
colunas = [coluna for coluna in dados.columns if dados[coluna].dtype == 'object']
codificador = preprocessing.LabelEncoder()
for coluna in colunas:
dados[coluna] = codificador.fit_transform(dados[coluna])
mascara = np.triu(dados.corr())
rfc = RandomForestClassifier(n_jobs=-1, n_estimators=500, max_depth=70, max_features=2, random_state=0)
knn = KNeighborsClassifier(n_neighbors=5, algorithm='kd_tree', weights='uniform', n_jobs=-1)
gbc = GradientBoostingClassifier(learning_rate=0.01, loss='exponential', max_depth=70, max_features=2, n_estimators=500, random_state=0)
rfc.fit(X_treino, y_treino)
# Para avaliar as métricas e a matriz de confusão, vamos criar uma função para facilitar a nossa vida
def metricas(X_teste, y_teste, classificador, nome_clf):
# Prever o resultado
y_pred = classificador.predict(X_teste)
# Métrica de acurácia
acuracia = accuracy_score(y_teste, y_pred)
# Métrica de precisão
precisao = precision_score(y_teste, y_pred)
# Métrica de revocação
revocacao = recall_score(y_teste, y_pred)
# Métrica de pontuação F1
f1 = f1_score(y_teste, y_pred)
# Mostrar os valores das métricas
print(f'''{nome_clf}
Acurácia: {acuracia:.3f}
Precisão: {precisao:.3f}
Revocação: {revocacao:.3f}
Pontuação F1: {f1:.3f}''')
# Criar a matriz de confusão
matriz = confusion_matrix(y_teste, y_pred)
# Criar um DataFrame para aramazenar os dados de "y_pred" e "y_teste"
df_mc = pd.DataFrame(matriz, index=['DCV', 'Normal'], columns=['DCV', 'Normal'])
# Determinar o tamanho da plotagem
plt.figure(figsize=(15, 10))
# Criar a plotagem
mapa_calor = sns.heatmap(data=df_mc,
annot=True,
cmap='Blues',
fmt='.5g',
annot_kws={'size': 20})
# Colocar o título da matriz de confusão
mapa_calor.set_title(f'Matriz de Confusão {nome_clf}', fontsize=25, y=1.05)
# Adicionar a legenda nos eixos
plt.xlabel('Valores Previstos', fontsize=20)
plt.ylabel('Valores Reais', fontsize=20)
# Mostrar a matriz de confusão
plt.show()
metricas(X_teste, y_teste, rfc, 'Random Forest Classifier') | code |
89141106/cell_48 | [
"image_output_1.png"
] | from matplotlib import style
from matplotlib.gridspec import GridSpec
import matplotlib.pyplot as plt
import numpy as np
import pandas as pd
import seaborn as sns
arquivo = '../input/heart-failure-prediction/heart.csv'
dados = pd.read_csv(arquivo)
dados
trocar_nomes = {'Age': 'Idade', 'Sex': 'Sexo', 'ChestPainType': 'Tipo de dor', 'RestingBP': 'Pressão', 'Cholesterol': 'Colesterol', 'FastingBS': 'Glicemia', 'RestingECG': 'Eletro', 'MaxHR': 'BPM', 'ExerciseAngina': 'Dor por exec.', 'ST_Slope': 'Incl. ST', 'HeartDisease': 'DCV'}
dados = dados.rename(columns=trocar_nomes)
dados
dados.describe().T
dados = dados[dados['Colesterol'] != 0]
dados = dados[dados['Pressão'] != 0]
dados.describe().T
sns.set_theme()
style.use('fivethirtyeight')
cores = ['lightcoral', 'deepskyblue', 'orchid', 'tomato', 'teal', 'darkcyan', 'limegreen', 'darkorange']
# Definir a função do gráfico de pizza
def grafico_pizza(data_frame, coluna, cores, explode, titulo, fonte):
# Fazer contagem dos valores da coluna selecionada
df = data_frame[coluna].value_counts()
# Determinar o tamannho da plotagem
plt.figure(figsize=(15, 10))
# Criar o gráfico de pizza
_, _, pacotes = plt.pie(df,
colors=cores,
labels=df.index,
explode=explode,
shadow=True,
startangle=90,
autopct='%1.1f%%',
textprops={'fontsize': fonte,
'color': 'black',
'weight': 'bold',
'family': 'serif'})
# Plotar o gráfico de pizza
plt.setp(pacotes, color='white')
# Colocar o título do gráfico
plt.title(titulo, size=45)
# Desenhar o círculo interno
circulo_centro = plt.Circle((0, 0), 0.40, fc='white')
fig = plt.gcf()
fig.gca().add_artist(circulo_centro)
# Definir o gráfico da função de distribuição
def grafico_distribuicao(data_frame, coluna, titulo):
# Armazenar os dados da coluna
dados = data_frame[coluna]
# Determinar a figura e seu tamanho
fig = plt.figure(figsize=(17, 7))
# Criar a grade em que os gráficos serão plotados
grade = GridSpec(nrows=2, ncols=1, figure=fig)
# Escolher uma das cores para o gráfico
cor = np.random.choice(cores, 1)[0]
# Motrar o valor de assimetria dos dados
print(f'Assimetria de {titulo}: {np.round(dados.skew(), 3)}')
# Plotar o histograma
ax0 = fig.add_subplot(grade[0, :])
ax0.set_title(f'Histograma e BoxPlot de {titulo}', y=1.05)
sns.histplot(data=dados, ax=ax0, color=cor)
# Plotar o BoxPlot
ax1 = fig.add_subplot(grade[1, :])
plt.axis('off')
sns.boxplot(x=dados, ax=ax1, color=cor)
grafico_pizza(dados, 'Sexo', ('#140E36', '#091AAB'), (0.05, 0.05), 'Sexo', 25) | code |
89141106/cell_41 | [
"text_plain_output_1.png",
"image_output_1.png"
] | from matplotlib import style
from matplotlib.gridspec import GridSpec
import matplotlib.pyplot as plt
import numpy as np
import pandas as pd
import seaborn as sns
arquivo = '../input/heart-failure-prediction/heart.csv'
dados = pd.read_csv(arquivo)
dados
trocar_nomes = {'Age': 'Idade', 'Sex': 'Sexo', 'ChestPainType': 'Tipo de dor', 'RestingBP': 'Pressão', 'Cholesterol': 'Colesterol', 'FastingBS': 'Glicemia', 'RestingECG': 'Eletro', 'MaxHR': 'BPM', 'ExerciseAngina': 'Dor por exec.', 'ST_Slope': 'Incl. ST', 'HeartDisease': 'DCV'}
dados = dados.rename(columns=trocar_nomes)
dados
dados.describe().T
dados = dados[dados['Colesterol'] != 0]
dados = dados[dados['Pressão'] != 0]
dados.describe().T
sns.set_theme()
style.use('fivethirtyeight')
cores = ['lightcoral', 'deepskyblue', 'orchid', 'tomato', 'teal', 'darkcyan', 'limegreen', 'darkorange']
# Definir a função do gráfico de pizza
def grafico_pizza(data_frame, coluna, cores, explode, titulo, fonte):
# Fazer contagem dos valores da coluna selecionada
df = data_frame[coluna].value_counts()
# Determinar o tamannho da plotagem
plt.figure(figsize=(15, 10))
# Criar o gráfico de pizza
_, _, pacotes = plt.pie(df,
colors=cores,
labels=df.index,
explode=explode,
shadow=True,
startangle=90,
autopct='%1.1f%%',
textprops={'fontsize': fonte,
'color': 'black',
'weight': 'bold',
'family': 'serif'})
# Plotar o gráfico de pizza
plt.setp(pacotes, color='white')
# Colocar o título do gráfico
plt.title(titulo, size=45)
# Desenhar o círculo interno
circulo_centro = plt.Circle((0, 0), 0.40, fc='white')
fig = plt.gcf()
fig.gca().add_artist(circulo_centro)
# Definir o gráfico da função de distribuição
def grafico_distribuicao(data_frame, coluna, titulo):
# Armazenar os dados da coluna
dados = data_frame[coluna]
# Determinar a figura e seu tamanho
fig = plt.figure(figsize=(17, 7))
# Criar a grade em que os gráficos serão plotados
grade = GridSpec(nrows=2, ncols=1, figure=fig)
# Escolher uma das cores para o gráfico
cor = np.random.choice(cores, 1)[0]
# Motrar o valor de assimetria dos dados
print(f'Assimetria de {titulo}: {np.round(dados.skew(), 3)}')
# Plotar o histograma
ax0 = fig.add_subplot(grade[0, :])
ax0.set_title(f'Histograma e BoxPlot de {titulo}', y=1.05)
sns.histplot(data=dados, ax=ax0, color=cor)
# Plotar o BoxPlot
ax1 = fig.add_subplot(grade[1, :])
plt.axis('off')
sns.boxplot(x=dados, ax=ax1, color=cor)
grafico_pizza(dados, 'DCV', ('#140E36', '#091AAB'), (0.05, 0.05), 'Doença Cardiovascular', 25) | code |
89141106/cell_54 | [
"image_output_1.png"
] | from matplotlib import style
from matplotlib.gridspec import GridSpec
import matplotlib.pyplot as plt
import numpy as np
import pandas as pd
import seaborn as sns
arquivo = '../input/heart-failure-prediction/heart.csv'
dados = pd.read_csv(arquivo)
dados
trocar_nomes = {'Age': 'Idade', 'Sex': 'Sexo', 'ChestPainType': 'Tipo de dor', 'RestingBP': 'Pressão', 'Cholesterol': 'Colesterol', 'FastingBS': 'Glicemia', 'RestingECG': 'Eletro', 'MaxHR': 'BPM', 'ExerciseAngina': 'Dor por exec.', 'ST_Slope': 'Incl. ST', 'HeartDisease': 'DCV'}
dados = dados.rename(columns=trocar_nomes)
dados
dados.describe().T
dados = dados[dados['Colesterol'] != 0]
dados = dados[dados['Pressão'] != 0]
dados.describe().T
sns.set_theme()
style.use('fivethirtyeight')
cores = ['lightcoral', 'deepskyblue', 'orchid', 'tomato', 'teal', 'darkcyan', 'limegreen', 'darkorange']
# Definir a função do gráfico de pizza
def grafico_pizza(data_frame, coluna, cores, explode, titulo, fonte):
# Fazer contagem dos valores da coluna selecionada
df = data_frame[coluna].value_counts()
# Determinar o tamannho da plotagem
plt.figure(figsize=(15, 10))
# Criar o gráfico de pizza
_, _, pacotes = plt.pie(df,
colors=cores,
labels=df.index,
explode=explode,
shadow=True,
startangle=90,
autopct='%1.1f%%',
textprops={'fontsize': fonte,
'color': 'black',
'weight': 'bold',
'family': 'serif'})
# Plotar o gráfico de pizza
plt.setp(pacotes, color='white')
# Colocar o título do gráfico
plt.title(titulo, size=45)
# Desenhar o círculo interno
circulo_centro = plt.Circle((0, 0), 0.40, fc='white')
fig = plt.gcf()
fig.gca().add_artist(circulo_centro)
# Definir o gráfico da função de distribuição
def grafico_distribuicao(data_frame, coluna, titulo):
# Armazenar os dados da coluna
dados = data_frame[coluna]
# Determinar a figura e seu tamanho
fig = plt.figure(figsize=(17, 7))
# Criar a grade em que os gráficos serão plotados
grade = GridSpec(nrows=2, ncols=1, figure=fig)
# Escolher uma das cores para o gráfico
cor = np.random.choice(cores, 1)[0]
# Motrar o valor de assimetria dos dados
print(f'Assimetria de {titulo}: {np.round(dados.skew(), 3)}')
# Plotar o histograma
ax0 = fig.add_subplot(grade[0, :])
ax0.set_title(f'Histograma e BoxPlot de {titulo}', y=1.05)
sns.histplot(data=dados, ax=ax0, color=cor)
# Plotar o BoxPlot
ax1 = fig.add_subplot(grade[1, :])
plt.axis('off')
sns.boxplot(x=dados, ax=ax1, color=cor)
grafico_distribuicao(dados, 'Pressão', 'Pressão arterial') | code |
89141106/cell_67 | [
"image_output_1.png"
] | from matplotlib import style
from matplotlib.gridspec import GridSpec
import matplotlib.pyplot as plt
import numpy as np
import pandas as pd
import seaborn as sns
arquivo = '../input/heart-failure-prediction/heart.csv'
dados = pd.read_csv(arquivo)
dados
trocar_nomes = {'Age': 'Idade', 'Sex': 'Sexo', 'ChestPainType': 'Tipo de dor', 'RestingBP': 'Pressão', 'Cholesterol': 'Colesterol', 'FastingBS': 'Glicemia', 'RestingECG': 'Eletro', 'MaxHR': 'BPM', 'ExerciseAngina': 'Dor por exec.', 'ST_Slope': 'Incl. ST', 'HeartDisease': 'DCV'}
dados = dados.rename(columns=trocar_nomes)
dados
dados.describe().T
dados = dados[dados['Colesterol'] != 0]
dados = dados[dados['Pressão'] != 0]
dados.describe().T
sns.set_theme()
style.use('fivethirtyeight')
cores = ['lightcoral', 'deepskyblue', 'orchid', 'tomato', 'teal', 'darkcyan', 'limegreen', 'darkorange']
# Definir a função do gráfico de pizza
def grafico_pizza(data_frame, coluna, cores, explode, titulo, fonte):
# Fazer contagem dos valores da coluna selecionada
df = data_frame[coluna].value_counts()
# Determinar o tamannho da plotagem
plt.figure(figsize=(15, 10))
# Criar o gráfico de pizza
_, _, pacotes = plt.pie(df,
colors=cores,
labels=df.index,
explode=explode,
shadow=True,
startangle=90,
autopct='%1.1f%%',
textprops={'fontsize': fonte,
'color': 'black',
'weight': 'bold',
'family': 'serif'})
# Plotar o gráfico de pizza
plt.setp(pacotes, color='white')
# Colocar o título do gráfico
plt.title(titulo, size=45)
# Desenhar o círculo interno
circulo_centro = plt.Circle((0, 0), 0.40, fc='white')
fig = plt.gcf()
fig.gca().add_artist(circulo_centro)
# Definir o gráfico da função de distribuição
def grafico_distribuicao(data_frame, coluna, titulo):
# Armazenar os dados da coluna
dados = data_frame[coluna]
# Determinar a figura e seu tamanho
fig = plt.figure(figsize=(17, 7))
# Criar a grade em que os gráficos serão plotados
grade = GridSpec(nrows=2, ncols=1, figure=fig)
# Escolher uma das cores para o gráfico
cor = np.random.choice(cores, 1)[0]
# Motrar o valor de assimetria dos dados
print(f'Assimetria de {titulo}: {np.round(dados.skew(), 3)}')
# Plotar o histograma
ax0 = fig.add_subplot(grade[0, :])
ax0.set_title(f'Histograma e BoxPlot de {titulo}', y=1.05)
sns.histplot(data=dados, ax=ax0, color=cor)
# Plotar o BoxPlot
ax1 = fig.add_subplot(grade[1, :])
plt.axis('off')
sns.boxplot(x=dados, ax=ax1, color=cor)
# Definir o gráfico de influência
def grafico_influencia(data_frame, coluna, bins, labels, com_bins=True):
# Armazenar os dados da coluna
influencia = data_frame.loc[:, [coluna, 'DCV']]
# Se os dados de "coluna" não forem como classe, então terá intervalos ("bins" e "labels")
if com_bins:
influencia[coluna] = pd.cut(influencia[coluna],
bins=bins,
labels=labels)
# Escolher uma das cores para o gráfico
cor = np.random.choice(cores, 1)[0]
# Determinar o tamanho da figura
plt.figure(figsize=(15, 5))
# Criar o gráfico
grafico = sns.pointplot(x=coluna, y='DCV', dodge=0.1, capsize=.1, data=influencia, color=cor)
# Colocar o título do gráfico
grafico.set_title(f'{coluna} influência', fontsize=25)
grafico_influencia(dados, 'Eletro', None, None, False) | code |
89141106/cell_50 | [
"image_output_1.png"
] | from matplotlib import style
from matplotlib.gridspec import GridSpec
import matplotlib.pyplot as plt
import numpy as np
import pandas as pd
import seaborn as sns
arquivo = '../input/heart-failure-prediction/heart.csv'
dados = pd.read_csv(arquivo)
dados
trocar_nomes = {'Age': 'Idade', 'Sex': 'Sexo', 'ChestPainType': 'Tipo de dor', 'RestingBP': 'Pressão', 'Cholesterol': 'Colesterol', 'FastingBS': 'Glicemia', 'RestingECG': 'Eletro', 'MaxHR': 'BPM', 'ExerciseAngina': 'Dor por exec.', 'ST_Slope': 'Incl. ST', 'HeartDisease': 'DCV'}
dados = dados.rename(columns=trocar_nomes)
dados
dados.describe().T
dados = dados[dados['Colesterol'] != 0]
dados = dados[dados['Pressão'] != 0]
dados.describe().T
sns.set_theme()
style.use('fivethirtyeight')
cores = ['lightcoral', 'deepskyblue', 'orchid', 'tomato', 'teal', 'darkcyan', 'limegreen', 'darkorange']
# Definir a função do gráfico de pizza
def grafico_pizza(data_frame, coluna, cores, explode, titulo, fonte):
# Fazer contagem dos valores da coluna selecionada
df = data_frame[coluna].value_counts()
# Determinar o tamannho da plotagem
plt.figure(figsize=(15, 10))
# Criar o gráfico de pizza
_, _, pacotes = plt.pie(df,
colors=cores,
labels=df.index,
explode=explode,
shadow=True,
startangle=90,
autopct='%1.1f%%',
textprops={'fontsize': fonte,
'color': 'black',
'weight': 'bold',
'family': 'serif'})
# Plotar o gráfico de pizza
plt.setp(pacotes, color='white')
# Colocar o título do gráfico
plt.title(titulo, size=45)
# Desenhar o círculo interno
circulo_centro = plt.Circle((0, 0), 0.40, fc='white')
fig = plt.gcf()
fig.gca().add_artist(circulo_centro)
# Definir o gráfico da função de distribuição
def grafico_distribuicao(data_frame, coluna, titulo):
# Armazenar os dados da coluna
dados = data_frame[coluna]
# Determinar a figura e seu tamanho
fig = plt.figure(figsize=(17, 7))
# Criar a grade em que os gráficos serão plotados
grade = GridSpec(nrows=2, ncols=1, figure=fig)
# Escolher uma das cores para o gráfico
cor = np.random.choice(cores, 1)[0]
# Motrar o valor de assimetria dos dados
print(f'Assimetria de {titulo}: {np.round(dados.skew(), 3)}')
# Plotar o histograma
ax0 = fig.add_subplot(grade[0, :])
ax0.set_title(f'Histograma e BoxPlot de {titulo}', y=1.05)
sns.histplot(data=dados, ax=ax0, color=cor)
# Plotar o BoxPlot
ax1 = fig.add_subplot(grade[1, :])
plt.axis('off')
sns.boxplot(x=dados, ax=ax1, color=cor)
grafico_pizza(dados, 'Tipo de dor', ('#5735FD', '#3C78E8', '#2E90FF', '#6186b0'), (0.05, 0.05, 0.05, 0.05), 'Tipo de dor', 17) | code |
89141106/cell_107 | [
"text_plain_output_1.png"
] | from sklearn.ensemble import GradientBoostingClassifier, RandomForestClassifier
from sklearn.neighbors import KNeighborsClassifier
rfc = RandomForestClassifier(n_jobs=-1, n_estimators=500, max_depth=70, max_features=2, random_state=0)
knn = KNeighborsClassifier(n_neighbors=5, algorithm='kd_tree', weights='uniform', n_jobs=-1)
gbc = GradientBoostingClassifier(learning_rate=0.01, loss='exponential', max_depth=70, max_features=2, n_estimators=500, random_state=0)
gbc.fit(X_treino, y_treino) | code |
89141106/cell_106 | [
"text_html_output_1.png"
] | from sklearn.ensemble import GradientBoostingClassifier, RandomForestClassifier
from sklearn.neighbors import KNeighborsClassifier
rfc = RandomForestClassifier(n_jobs=-1, n_estimators=500, max_depth=70, max_features=2, random_state=0)
knn = KNeighborsClassifier(n_neighbors=5, algorithm='kd_tree', weights='uniform', n_jobs=-1)
gbc = GradientBoostingClassifier(learning_rate=0.01, loss='exponential', max_depth=70, max_features=2, n_estimators=500, random_state=0)
knn.fit(X_treino, y_treino) | code |
89141106/cell_45 | [
"text_plain_output_1.png",
"image_output_1.png"
] | from matplotlib import style
from matplotlib.gridspec import GridSpec
import matplotlib.pyplot as plt
import numpy as np
import pandas as pd
import seaborn as sns
arquivo = '../input/heart-failure-prediction/heart.csv'
dados = pd.read_csv(arquivo)
dados
trocar_nomes = {'Age': 'Idade', 'Sex': 'Sexo', 'ChestPainType': 'Tipo de dor', 'RestingBP': 'Pressão', 'Cholesterol': 'Colesterol', 'FastingBS': 'Glicemia', 'RestingECG': 'Eletro', 'MaxHR': 'BPM', 'ExerciseAngina': 'Dor por exec.', 'ST_Slope': 'Incl. ST', 'HeartDisease': 'DCV'}
dados = dados.rename(columns=trocar_nomes)
dados
dados.describe().T
dados = dados[dados['Colesterol'] != 0]
dados = dados[dados['Pressão'] != 0]
dados.describe().T
sns.set_theme()
style.use('fivethirtyeight')
cores = ['lightcoral', 'deepskyblue', 'orchid', 'tomato', 'teal', 'darkcyan', 'limegreen', 'darkorange']
# Definir a função do gráfico de pizza
def grafico_pizza(data_frame, coluna, cores, explode, titulo, fonte):
# Fazer contagem dos valores da coluna selecionada
df = data_frame[coluna].value_counts()
# Determinar o tamannho da plotagem
plt.figure(figsize=(15, 10))
# Criar o gráfico de pizza
_, _, pacotes = plt.pie(df,
colors=cores,
labels=df.index,
explode=explode,
shadow=True,
startangle=90,
autopct='%1.1f%%',
textprops={'fontsize': fonte,
'color': 'black',
'weight': 'bold',
'family': 'serif'})
# Plotar o gráfico de pizza
plt.setp(pacotes, color='white')
# Colocar o título do gráfico
plt.title(titulo, size=45)
# Desenhar o círculo interno
circulo_centro = plt.Circle((0, 0), 0.40, fc='white')
fig = plt.gcf()
fig.gca().add_artist(circulo_centro)
# Definir o gráfico da função de distribuição
def grafico_distribuicao(data_frame, coluna, titulo):
# Armazenar os dados da coluna
dados = data_frame[coluna]
# Determinar a figura e seu tamanho
fig = plt.figure(figsize=(17, 7))
# Criar a grade em que os gráficos serão plotados
grade = GridSpec(nrows=2, ncols=1, figure=fig)
# Escolher uma das cores para o gráfico
cor = np.random.choice(cores, 1)[0]
# Motrar o valor de assimetria dos dados
print(f'Assimetria de {titulo}: {np.round(dados.skew(), 3)}')
# Plotar o histograma
ax0 = fig.add_subplot(grade[0, :])
ax0.set_title(f'Histograma e BoxPlot de {titulo}', y=1.05)
sns.histplot(data=dados, ax=ax0, color=cor)
# Plotar o BoxPlot
ax1 = fig.add_subplot(grade[1, :])
plt.axis('off')
sns.boxplot(x=dados, ax=ax1, color=cor)
# Definir o gráfico de influência
def grafico_influencia(data_frame, coluna, bins, labels, com_bins=True):
# Armazenar os dados da coluna
influencia = data_frame.loc[:, [coluna, 'DCV']]
# Se os dados de "coluna" não forem como classe, então terá intervalos ("bins" e "labels")
if com_bins:
influencia[coluna] = pd.cut(influencia[coluna],
bins=bins,
labels=labels)
# Escolher uma das cores para o gráfico
cor = np.random.choice(cores, 1)[0]
# Determinar o tamanho da figura
plt.figure(figsize=(15, 5))
# Criar o gráfico
grafico = sns.pointplot(x=coluna, y='DCV', dodge=0.1, capsize=.1, data=influencia, color=cor)
# Colocar o título do gráfico
grafico.set_title(f'{coluna} influência', fontsize=25)
grafico_influencia(dados, 'Idade', [0, 30, 40, 50, 60, 70, 100], ['<30', '30-40', '40-50', '50-60', '60-70', '70+']) | code |
89141106/cell_18 | [
"image_output_1.png"
] | import pandas as pd
arquivo = '../input/heart-failure-prediction/heart.csv'
dados = pd.read_csv(arquivo)
dados
trocar_nomes = {'Age': 'Idade', 'Sex': 'Sexo', 'ChestPainType': 'Tipo de dor', 'RestingBP': 'Pressão', 'Cholesterol': 'Colesterol', 'FastingBS': 'Glicemia', 'RestingECG': 'Eletro', 'MaxHR': 'BPM', 'ExerciseAngina': 'Dor por exec.', 'ST_Slope': 'Incl. ST', 'HeartDisease': 'DCV'}
dados = dados.rename(columns=trocar_nomes)
dados
dados.info() | code |
Subsets and Splits
No saved queries yet
Save your SQL queries to embed, download, and access them later. Queries will appear here once saved.