Phishing Site Detection Classification Model Using Machine Learning Approach

Abstract

Phishing has been a cybercrime that has existed for a long time, and there are still many people who are victims of this attack. This research attempts to prevent phishing by extracting the attributes found on phishing websites. This study uses a hybrid method by combining allowlist and denylist as part of a classification system. This research utilizes 18 features to identify a phishing site in terms of address bar, abnormal request, and source code (HTML and JavaScript). Where in each feature the author determines the benchmark. This study validates the status code and detects 52 URL shortening service domains and then evaluates these abnormalities with a binary classification system. Algorithms that have good results are Decision Tree and K Nearest Neighbor (KNN). After evaluating the performance of the algorithm in terms of Precision, Recall, and F-Measure. As a result, the Decision Tree algorithm has the highest accuracy of 97.62% and the fastest computation time of 0.00894 seconds. So that the Decision Tree is superior in terms of accuracy and computation time in detecting phishing URLs.