Phish-blocker: Safeguarding against Phishing Attacks in Real-Time

Phishing is defined as a cyber attack that uses social engineering through digital tools to persuade victims to reveal personal information such as their passwords or credit card numbers. Finally, stolen personal information is used to fraudulently trust legitimate websites or financial institutions and generate illegal profits. Although there are different solutions against phishing, phishing attacks have increased significantly in the last few years. Some solutions rely on removing features by rules, while some features rely on third-party services, which can cause instability and prolong problems with prediction services. This project introduced Phish Wiper, a deep learning framework that uses neural networks (RNN) to identify URL and HTML features in a real-time search engine. Phish Wiper uses two deep encryption methods (URL blocks and HTML blocks) that are learned separately and together throughout the connection process, eliminating each network's publishing process as the final decision. This method examines the URL and HTML of web pages and calculates their similarity to known phishing websites to identify them. Phishing detection is a dual task; There are two categories: legitimate and phishing. We use this framework as a browser plug-in that can instantly detect and alert when a user visits a web page if there is a phishing risk. According to the experimental results, the proposed model achieved good performance with an accuracy of 96.79% to 98.90% when measured using different data.