Static code analysis and detection of multiple malicious Java applets using SVM

An applet that performs an action against the will of the user who invoked it should be considered malicious. A malicious applet is applet that attacks the local system of a Web surfer. They can even seriously damage a Java user´s machine. The problem of malicious Java applets, that is currently not well addressed by existing work. We have developed a tool for malicious Java applets, which we call Jarhead. The approach is based on static code analysis. The approach extracts features from Java applets, and uses machine learning technique called support vector machine(SVM) to produce a tool. This approach is able to detect both known and previously-unseen real-world malicious applets.