Enhancing Phishing Detection Through Advanced Machine Learning Techniques

With the rapid growth of internet usage, phishing attacks have become increasingly prevalent, presenting a major challenge to cybersecurity. This paper introduces a sophisticated method for detecting phishing websites by leveraging advanced machine learning algorithms. Specifically, Evaluated the performance of Logistic Regression, Linear Regression, and Multinomial Naive Bayes models in distinguishing phishing websites based on their URL features. The study utilizes a robust dataset comprising over 100,000 URLs, each meticulously classified as either legitimate or phishing. Through comprehensive data preprocessing and an innovative optimization pipeline, Refined the Logistic Regression model to achieve a detection accuracy of 98.0369%. Additionally, incorporated the web parameter analysis to enhance detection precision and minimize false positives. This work significantly contributes to the development of more resilient cybersecurity strategies, addressing key issues such as feature selection, handling large datasets, and adapting to the constantly evolving nature of phishing tactics.