Abstract
Phishing has grown into one of the major and supreme operative in cyber threats, triggering millions of data breaches and security failures every year. This paper proposes a CNN based prediction analysis using Optimistic Multi centric feature extraction for phishing attack detection technique that uses only URL functions to expedite and accurately locate phishing websites and explore structured databases. Using anti-phishing technology requires experts to extract the characteristics of phishing URL sites and use behavioral principle which is identified through URL Behavioral Rectifier (U-BR). Then the feature dependability is supported to create URL probability phishing index (U-PFI) to identify the relative weight for detection of phishing sites. By attaining the feature weight, the features are observed using Optimized Multi Centric Feature Selection (OMCFS) issued to reduce the dimension Log variation, and then the selected features get trained through Conventional Neural Network (CNN). This method predicts the legitimacy of URLs without the access to web content to find the phishing and depends on domain related service. The proposed technique converts URLs into standard size scales using writing embedding techniques, separates features at different levels using the CNN model, and classifies the features as risk by category.
Shitharth Selvarajan
Lecturer in Cyber Security
My research interests include Cyber Security, Blockchain, Critical Infrastructure & Systems, Network Security & Ethical Hacking.