Adaptive Firewall System to Predict Phishing Websites using Machine Learning Model

With an increase in occurrence of cyber threats and phishing attacks that pose a threat to network security, this paper presents a way in identifying and blocking these threats with the help of a machine learning model which can detect and block phishing websites. By analyzing website URLs and their features, and generates firewall rules. The adaptive firewall is integrated with user-friendly Streamlit interface which helps identifying the legitimacy of an URL for real time rule management and monitoring. The system in identifying phishing sites has an accuracy of 97%. The proposed methodology uses a Random Forest Classifier for the easy detection of phishing websites and uses Uncomplicated Firewall (UFW) for providing rules. The model was trained for about 11,000 URLs which help in classifying and Streamlit provides an interactive interface to visualize the results obtained by the model. The research is done by combing computer networks and machine learning for an intelligent adaptive firewall system. It offers a solution that can be deployed in various environments example educational institutions.