Abstract-This paper presents a review of machine learning-based approaches for identifying and classifying malicious websites. Various studies have tackled this challenge by extracting features from URLs and website content to train machine learning models. Methods include lexical, host-based, and content-based feature extraction, as well as domain and Alexa-based analyses. Different algorithms, including Gradient Boosting, Random Forests, and Neural Networks, have been utilized, achieving high accuracy in classifying malicious URLs. Real-time concept drift detection and retraining models have also been proposed to adapt to evolving cyber threats. The paper concludes by highlighting the importance of feature extraction in bolstering cybersecurity.
