PHIKITA: PHISHING KIT ATTACKS DATASET FOR PHISHING WEBSITES IDENTIFICATION

Abstract

According to recent research, phishers are making use of phishing kits to launch increasingly frequent, extensive, and automated phishing assaults. One possible way to catch phishing efforts early is to check deployed websites for phishing kits. Our research has not found any databases that provide a collection of phishing kits utilised by malicious websites. Here, we provide PhiKitA, an innovative dataset that includes both phishing kits and the websites that are built utilising them. In three tests, we have used graph representation DOM techniques, MD5 hashes, and fingerprints to get baseline findings in PhiKitA: phishing website identification, identifying the source of a phishing website, and familiarity analysis of phishing kit samples. The familiarity analysis reveals a little phishing effort and many phishing kits. The graph representation technique attained a 92.50% accuracy rate in the binary classification
issue for phishing detection, indicating that the data included in phishing kits contains valuable information for phishing classification. Lastly, the MD5 hash representation produced an F1 score of 39.54%, indicating that this approach is unable to adequately extract sufficient information to differentiate between phishing websites and the suppliers of their phishing kits.