An Improved Evolution Algorithm of Immune Detectors for Network Data Analysis

Abstract

Traditional immune algorithms use binary strings to represent detectors and adopt r-contiguous matching algorithm to match detectors. It reduces the accuracy of network data analysis. In order to raise the above performance of network data analysis based on immune algorithms, an improved evolution algorithm of immune detectors for network data analysis is proposed in this paper. Traditional creation method, traditional dynamic evolution method and traditional matching method are analyzed. Network data are simulated with network packets. Immune detectors are simulated. Computation algorithm of similarity is set up. Generation algorithm of immune detector is designed. Based on the above simulation and sub algorithms, the total network data analysis algorithm is constructed. A prototype software is developed to verify the effectiveness of the proposed algorithm. The experiment results show that the proposed immune algorithm has better performance.