k-code/fraud-detection-model

SMS Spam Classifier

This application uses a bidirectional LSTM model to classify SMS messages as either spam or legitimate (ham). Simply enter your text message, and the model will predict whether it's spam or not, along with a confidence score.

Usage

1. Enter your text message in the input box
2. Click submit
3. The model will return its prediction (spam/ham) and confidence level

Model

The classifier uses a bidirectional LSTM architecture with:

• Word embeddings
• 2 LSTM layers
• Dropout for regularization
• Dense layers with ReLU activation