Abstract

In today's world, the use of websites for purposes like e-commerce and entertainment has become extremely common. However, one significant concern is determining whether a website is safe or fraudulent. Traditionally, browsers offer protection services that flag harmful websites, especially those redirecting to malicious sites, by marking them with a warning symbol. Despite these measures, browsers often fail to detect phishing websites, which are not inherently malicious but instead deceive users into providing sensitive information unknowingly. Our project addresses this issue by developing a machine learning model to identify phishing sites based on URL features such as domain length and character composition. We will train the model using various algorithms, evaluating each one's performance. By comparing the results, we aim to select the most accurate algorithm to reliably detect phishing sites. This approach ensures better protection for users against data theft and enhances overall internet safety.