The Research Of Intrusion Detection Based On Intelligence

Today's age is an Information age, and the internet is inseparable to people'life. The internet brings convenience for people, at the same time it also brings its own security problem. Nowadays people take these safety technologies, such as Router filtering, Firewall Intercept, Blocking Loophole and so on .When unknown internet attacks occur, these static security technologies are very hard to resist effectively. Intrusion Detection is a kind of active defense technology, and makes up traditional passive technology weaknesses. Computer Intrusion Real-Time Detection System could monitor the hosts and network strictly, analyze the suspicious events .When the Network Intrusion is detected, the system will take appropriate measures to eliminate the potential hazards in time and to ensure the security of the system.Aiming at missing report rate and false alert rate which exist generally in the Intrusion detection system , this paper discusses a intelligentized and intrusion detection model .Based on the characteristics of global superiority of genetic algorithm and locality of nerve, the model optimizes the weights of the neural network Using genetic algorithm. Experiment results show that the intelligentized way can improve the efficiency of the intrusion detection.This paper studies the main contents as follows:(1) Introducing the concept of the intrusion detection and flow characteristics, Comparing recognition differences in the neural network under different optimization method.(2) Analysising and researching the genetic algorithm. According to the defects of easily precocious of genetic algorithm and poor local search ability, analysising some the optimization strategy for genetic algorithm.(3) Proceeding intrusion detection through intelligentized ways.Experiments show that it can to a certain extent enhance inspection accuracy.