An Intelligent Hybrid ML System for URL-Based Phishing Threat Detection

Abstract: This project utilizes machine learning techniques to detect phishing URLs by analyzing various features of both legitimate and fraudulent websites. With the rapid growth of internet usage and online transactions such as banking, shopping, and communication, phishing attacks have become one of the most serious cybersecurity threats. Attackers create deceptive websites that closely resemble legitimate ones in order to trick users into revealing sensitive information such as usernames, passwords, and financial details.To address this issue, the proposed system uses a Multilayer Perceptron (MLP) algorithm, which is a type of artificial neural network capable of learning complex patterns from data. The system is designed to classify URLs as either phishing or legitimate based on a set of extracted features. These features include URL length, number of special characters, number of subdomains, use of HTTPS protocol, and other structural characteristics that help differentiate between safe and malicious websites.The system is trained using a dataset that contains a large number of URLs labeled as either phishing or legitimate. During the training phase, the model learns the patterns and relationships between the extracted features and the corresponding labels. This enables the system to understand the common characteristics of phishing URLs and distinguish them from legitimate ones.Once the training process is completed, the model is ready for real-time prediction. When a user inputs a URL into the system, it undergoes the same feature extraction process. The extracted features are then passed to the trained MLP model, which analyzes the data and predicts whether the URL is safe or a phishing attempt. The result is then displayed to the user, helping them avoid accessing malicious websites.The proposed system plays an important role in enhancing online security by preventing unauthorized access to sensitive information. It provides a fast and efficient way to detect phishing attempts and reduces the risk of cyber-attacks. Compared to traditional methods, this machine learning-based approach is more flexible and capable of identifying new and previously unseen phishing URLs. Keywords- Phishing Detection, Machine Learning, URL Analysis, Cybersecurity, Multilayer Perceptron (MLP), Random Forest, Feature Extraction, Classification, Data Security, Hybrid Model