PHISHING DETECTION AND PREVENTION USING MACHINE LEARNING ALGORITHMS

ABSTRACT The Internet has become an integral part of our lives. With the increased use of the internet, fraudsters have become more active and advanced. Phishing attacks are one of the most serious security risks on the Internet today. Due to technological advancements, detecting phishing attempts has become more challenging. Understanding if a web page is legitimate or not is a difficult task because of its semantics-based attack strategy, which primarily exploits computer users' vulnerabilities. The goal of this research is to create a machine learning model that can enhance accuracy and reduce false positives when detecting phishing emails and URLs. There are several methods for determining whether a website is legitimate or not. Many software companies are launching anti-phishing technologies that use approaches such as signatures, blacklists, heuristics, and visuals. However, each has its advantages and disadvantages. Therefore, there is a need for real-time machine learning methods. Using algorithms such as random forests, xgboost and neural networks for the classification of phishing webpages, we can achieve our goal. To increase detection accuracy, we use advanced feature extraction techniques such as URL Length, Web Traffic, Page Rank, and HTTPS token. Data will be collected from publicly available Kaggle databases. Model performance will be assessed using accuracy, precision, and recall. This study not only explores how we may use several classification models to improve the accuracy of detecting phishing websites but also addresses the practical implementation challenges associated with integrating these models into existing security infrastructures. Keywords- Phishing detection, machine learning, cybersecurity, data analysis, classification