Research on network malicious code dendritic cell immune algorithm based on fuzzy weighted support vector machine

Abstract

According to the significant impact on the accuracy rate of detection of current immune algorithms brought by incorrect classification of signal, it proposes network malicious code dendritic cell immune algorithm based on fuzzy weighted support vector machine. It summarizes and compares the pros and cons of the four methods, which are basic artificial intelligence immune algorithm, detection of unknown viruses based on immunity theory, malicious code immunity based on cryptography and automated intrusion response based on danger theory. It elaborates the process of the algorithm of the proposed immune algorithm, discusses a variety of input and output signals, gives the samples of the actual input signals, applies the coefficient of variation method to determine the values of the weights so as to enhance the discrimination ability of signal processing results, describes the principle and algorithm steps of fuzzy weighted support vector machine clustering method within immune algorithm proposed. By comparing the fuzzy aggregation before and after the immunization program experiments, it draws the conclusion that the proposed immune algorithm can optimize the input signal, fuzzy clustering the signal and the antigen, so as to bring down the number of immunization strategies and reduce the immune response time, as a result it improve the efficiency and performance of the immune system. © 2013 Springer-Verlag.