User Behavioral Characteristics Mining Algorithms Based on Net-Flow

Abstract

Network security technology-based firewall, intrusion detection, less user behavior from the network point of view to is considered possible vandalism. Net Flow collector for this problem, the network traffic statistics, the design of said data structures and statistical models of user behavior characteristics, the behavior of specific information, the establishment of the database behavior in a certain period of time granularity; on the basis of the behavior of database the design the clustering mining algorithm applicable to user behavior features, defined from the user behavior, determine the behavior patterns of individual users of the network. The experiments show that the design algorithm can effectively tap the behavior of the user's network, and provide a valid basis for management, analysis of user behavior.