🔗 MaliceSpotter

# Task

Develop a system that can accurately classify URLs as phishing or legitimate to prevent cyber attacks and protect users from malicious online activity.

# Solution

MaliceSpotter is a cybersecurity system that detects phishing URLs using an ensemble machine learning approach. Built with a Flask frontend and Python backend, it analyzes URLs based on 28 parameters, including the presence of an IP address, URL length, domain age, HTTPS usage, and suspicious redirection patterns.

The system employs three machine learning algorithms: Logistic Regression, Random Forest, and K-Nearest Neighbors (KNN), all combined through a hard Voting Classifier. Logistic Regression processes multiple parameters via the Sigmoid function, Random Forest builds decision trees using the Gini criterion, and KNN classifies URLs based on feature similarity and the majority class of nearest neighbors. The Voting Classifier determines the final prediction based on majority vote, ensuring robust decision-making.

This ensemble approach enhances accuracy, achieving 95% overall, outperforming individual models: Logistic Regression (92%), Random Forest (96%), and KNN (94%). Combining multiple models thus produces a more reliable phishing detection system than single-algorithm methods.

Research Work: https://indjst.org/articles/malicespotter-revolutionizing-cyber-security-with-machine-learning-for-phishing-resilience