A METHOD FOR THE ATTRIBUTE-BASED DETECTION OF URLS USING FREQUENCY PATTERNS

Abstract

In this paper a new heuristic method of URLs detection using host-based attributes called Frequency Patterns Analysis (FPA) is discussed. A detection criterion is probabilistic and is based on frequency patterns of host-based attributes specific to a certain class of URLs: benign, malicious or phishing. In such a way, users need to have the patterns of host-based attributes to detect a malicious web resource. The objective of the study is to mine frequency patterns of host-based attributes and use them to identify cyber-attacks on an individual cyberspace (IC) . A link (association) analysis which is a part of data mining methodology was used to discover frequency patterns.