Modeling User Interests Based on Cloud Model for Masquerade Detection

Masquerade detection techniques most commonly used are based on the premise that the behaviors the masquerader performs are inconsistent with what the victim performs. However, they fail when the attacker mimics the legitimate user's behaviors to a large extent. The attacker, in fact, has his own unique interests for the resources he gains access to by masquerading a legitimate user. For this reason, to learn the user specific interests will help in identifying the masqueraders. This is a new attempt. Based on the outstanding characteristics of cloud model on processing transformation between a qualitative concept and a set of quantitative numerical values, we apply cloud model to analyzing the correlation between user behaviors and user interests. And then we use the result to detect masquerades. The experiment verifies the feasibility of our method.