Anomaly Detection for User Behavior in Wireless Network Based on Cross Entropy

Wi-Fi technology is becoming more and more popular and anomaly detection for user behavior is much required in many fields such as public security, therefore anomaly detection for offline user behavior based on Wi-Fi is attracting many researchers´ attention. There are already some researches in anomaly detection focusing on single user behavior, while the distribution of group user behaviors is considered in this paper. A novel method using cross entropy model which is usually used for networking attacking in computer network is proposed to detect anomalies of offline user behavior in wireless network. Moreover, to make the anomaly detection available in different scenes, cross entropy model is optimized by using membership function of fuzzy mathematic. Experiments conducted in campus, hotel and office satisfy the expected results, which verify the feasibility and efficiency. And the result retrieved from optimized model is better than that of original model.