Enhancing Malicious URL Detection with Advanced Machine Learning Techniques

Malicious URLs pose a significant threat to online security. The can steal our personal information and can even hack out system. Refraining ourselves from using such malicious websites is the need of the hour. This paper presents a comprehensive evaluation of various machine learning algorithms which are used to effectively detect and mitigate these threats. This research explores the performance of traditional machine learning techniques including Random Forests,Decision Trees, Logistic Regression, Support Vector Machines, as well as advanced techniques like XGBoost, LightGBM, and CatBoost while the previous works focused on only traditional classification techniques. Additionally, this work investigates the potential of deep learning models, including LSTM and CNN, to capture intricate patterns within URL sequences. Machine learning algorithms are evaluated using the evaluations metrics including Accuracy, Recall, Precision and F1 score. Experimental Results and analysis have shown that advanced models outperform traditional classification techniques. These results valuable insights into the most effective approaches for safeguarding online users from malicious attacks.