PHISHING DETECTION AND SYSTEM AWARENESS USING MACHINE LEARNING

Abstract

As a large portion of our financial transactions, professional tasks, and everyday activities have shifted to the online environment, the exposure to cyber threats has significantly increased. Among these threats, URL-based phishing attacks are one of the most prevalent dangers faced by internet users. In such attacks, cybercriminals take advantage of human weaknesses rather than technical system vulnerabilities. These attacks are directed at both individuals and organizations, encouraging users to click on seemingly trustworthy links in order to obtain sensitive information or introduce malicious software into their systems. To combat this issue, various machine learning techniques have been developed to detect phishing URLs by classifying them as either malicious or legitimate. Researchers are continuously working to enhance the effectiveness of these models and improve their accuracy. This study focuses on reviewing different machine learning approaches used for phishing URL detection, along with the datasets and URL-based features utilized in training these models. Furthermore, it examines and compares the performance of multiple algorithms and discusses strategies employed to improve their accuracy. The objective is to provide a comprehensive survey that helps researchers understand recent advancements in this domain and supports the development of more reliable and precise phishing detection systems