Research On Real-valued Negative Selection Algorithm Based On Information Entropy Theory

The rapid development of Internet technology has brought great convenience to people’s production and life. With the full enjoy the convenience brought by the network people also can not ignore the issue of network security. Today, traditional network security technology has been difficult to meet the requirements of the existing network. The intrusion detection system based on artificial immune which is a network security means having active defense capabilities gets more and more the concern of experts and scholars This dissertation takes negative selection algorithm that is the core algorithm of the intrusion detection system based on artificial immune as research object, focuses on the theme of real-valued negative selection algorithm performance, and discuesses the methods of improving performance of real-valued negative selection algorithm.Intrusion detection system based on artificial immune detects the attack by the detector. The real-valued detector in the training phase and testing phase need to extract data in the network packet that constitutes a packet attribute vector., All the components of the attribute vector extracted is not very useful, and you need to select these property characteristics. To this issue, this paper proposes a feature selection algorithm of intrusion detection system based on immune.In this algorithm information entropy theory is applied to the attribute feature selection. Information entropy can be used to describe the size of the information content of the random variable.The application of information entropy theory in the detector feature selection is a good performance of the global measure and we can find characteristics that contain the most information with it. The experimental results show that this method improves the performance of the intrusion detection system based on artificial immune due to improve the efficiency of feature selection. The detection efficiency of the real value detector in high-dimensional shape space is low. In order to improve the detection of real-valued detectors in high-dimensional shape space, to reduce the dimension of shape space can be used. To this issue, this paper proposes a real-valued negative selection algorithm based on entropy-weighted. This algorithm works by the information content to determine the importance of each attribute of the space and selects the features that have more information content to reconstruct the the space to complete the form conversion from low-dimensional space to the high-dimensional. The traditional affinity calculated using squared Euclidean distance is unreasonable. Matching rules used by the algorithm proposed in this chapter is a weighted Euclidean distance, given different weights according to the degree of importance of each attribute in the calculation of affinity, attribute weights were chosen from the entropy of corresponding properties, so affinity of the calculation is more accurate. Experimental results show that this algorithm can supply the deficiency of real-valued detector generation algorithm in high dimensional space, and improve the detection performance of detector in high dimensional space.