Development of comprehensive taught system that protects information resources from phishing attacks

This research is devoted to development of comprehensive taught system that protects information resources from phishing attacks. Based on thorough analysis of available phishing resources, a set of identifiers, characteristics of phishing websites, are determined. The possibility of detecting these identifiers has been studied, and corresponding algorithms have been developed. The main challenge of this research was to analyse the cumulative effect of identifiers. It was addressed during system development stage. As a result, a mechanism (based on optimization techniques) that determines the risk level of a certain website has been developed. Learning capability of the system is based on data mining. Specifically, neural network technology and linear regression methods have been used extensively. Existing phishing websites databases have been used to create a learning sample.