A binary classification model to detect malicious URLs using a Bidirectional LSTM network. In response to the prevalent cyber-threat of phishing attacks, this project applies deep learning to natural language processing (NLP) to effectively detect potential phishing attacks.
Install python and the required libraries (must have pip installed). Ensure that the repository is downloaded and the command is run in the appropriate directory:
$ pip install -r install-libraries.txt
Once the model is downloaded (updated .h5 model will be uploaded soon), run the following:
$ python flask_rest_api.py
Next, open a new tab or window and run the following, replacing https://phishtank.org/faq.php with a URL (non-base URL) of your choice:
$ python3 flask_request.py -u https://www.https://phishtank.org/faq.php
Expected Output:
$ [{'Prediction Result': 3.228831036385466, 'Malicious URL Probability': 'The URL provided is not likely to be malicious.', 'url': ' https://phishtank.org/faq.php'}]