Random Forest Phishing Detector
Model Description
This is a Random Forest classifier trained to detect phishing emails. It utilizes a combination of TF-IDF text vectorization on the email text and structural/heuristic features (such as hidden HTML tags, URL anomalies, and domain mismatches).
Intended Use
This model is designed to evaluate the probability that an incoming email is a phishing attempt. It serves as a core component of a larger machine learning ensemble system.
Performance
Based on the test set evaluation:
- Accuracy: ~96%
- Precision: ~97%
- Recall: ~94%
- F1-Score: ~95%
Note: This model is saved as a scikit-learn .pkl file. Because it is not a deep learning model (like PyTorch or TensorFlow), Hugging Face's automated Inference API widget on this page will say it is 'not deployed'. This is completely normal and the model can still be downloaded and run locally.
