Proper

Abstract

Misuse detection method and anomaly detection method are widely used for the detection of malicious web pages. Both are based on machine learning. Misuse detection can detect known malicious web pages, but it cannot detect new ones. In contrast, anomaly detection can detect unknown malicious web pages, but it has a high false positive rate. In order to achieve a high detection rate through precisely detecting known and unknown malicious web pages, we propose a two-phase detection scheme. In the first phase, the misuse detection model is built based on the C4.5 decision tree algorithm, which allows known malicious web pages to be detected. In the second phase, the anomaly detection model with a one-class support vector machine is used to detect new types of malicious web pages. The experimental results show that our proposed method has significantly higher malicious web page detection rate than conventional ones with the expense of slightly higher false positive rate.