Research And Application Of Intrusion Detection Technology Based On Feature Analysis And Support Vector Machines

No matter what kind of network security requirements,intrusion detection technology is an important part of network security.Discovering the intrusions in the concurrent network discovery is very important,and these are the premise of making security policy.So for,efficient intrusion detection technology is necessary in intrusion detection system.According to the correlation analysis of intrusion behavior and the intention of attacker's invasion,it can deduce the potential attacks by intrusion detection technology.Therefore,the effective intrusion detection technology can not only provide the reliable basis for the development of effective defense measures,but also strive for precious time for the system.With the rapid development of the Internet and in the large scale network environment,traditional intrusion detection technology can not meet the demand of the detection accuracy and detection efficiency.Therefore,intelligent detection technology has an important significance in the rapidly growing amount of data of the network environment.In order to improve the detection accuracy,reduce the false positive and false negative and improve the detection efficiency,the dissertation discusses the feature analysis and support vector machines technology,then establishes a new type of intelligent intrusion detection model.The proposed classification model is applied to the intrusion detection area and the application of the classification model algorithm is specific introduced in the area.Then achieve the specific design and implementation of the key modules of the system.The dissertation mainly induces the following work.1.The problems of the intrusion detection on the internet is analyzed.Then the development course of domestic and international,research production,research trends of the dissertation is summarized in this dissertation.2.The relevant theoretical knowledge used in the dissertation is summarized,including support vector machines,feature selection methods,parameters optimization of support vector machine.These are foreshadowing for the following work.3.Pointing at the disadvantages of exiting intrusion detection technology,an optimization algorithm which combines artificial bee colony algorithm and support vector machine is proposed.According to the characteristics of intrusion detection,the dissertation analyze the astringency of artificial bee colony algorithm.It is used to do the feature selection with the data and design a new nectar individual,fitness function.At the same time,it combines the feature selection and SVM classification model for parameter optimization.The established classification model can ensure the accuracy of intrusion detection and improve the efficiency of detection.4.Introduce the application of the proposed optimization algorithm in the area of intrusion.Based on the algorithm,an classification model is established.The dissertation introduce the frame of the system and the simulation is carried out on this classification model.5.The intrusion detection classification model is tested,and the performance of the model is analyzed.