Research And Implementation On Multi-source Information Fusion Algorithm For Intrusion Detection

In the era of big data,data fusion technology has become an indispensable science and technology.In an intrusion detection system,the information collected by sensors must contain redundant information.In many cases,the information is conflicting,but there are also interrelated parts.The cooperation of these information determines the generation of a decision.How to deal with the information of conflict and cooperation is the key problem we need to consider in the fusion module.This paper mainly studies the multi-source information fusion algorithm for intrusion detection.(1)First proposed a multi-source data fusion conflict solution,the method adopts the error distribution to calculate the weights of data source,can accurately reflect the reliability of the data source,the experimental results show that the method can well solve the problem of multi-source conflict,this method shows its benchmark compared to other methods to improve the prediction accuracy,and on the two data sets are displayed MNAD(Mean Normalized Absolute Distance)and RMSN(Root Mean Squared Error)superiority,and a slight increase in speed.(2)Then,the classical data fusion method based on d-s evidence theory is improved,and the overall weight ratio of each evidence body is calculated and allocated mainly by introducing Pearson correlation coefficient.And for the shortcomings of the synthesis rule,the synthesis rule is improved to obtain a new fusion algorithm.Experiments show that the fusion value of the algorithm is higher than that of the benchmark method for the basic probability of reasonable proposition,and it has good fusion effect and fast convergence speed.As the amount of evidence increases,the superiority becomes more apparent.(3)Finally,an intrusion detection system based on data fusion is designed and implemented,which is divided into four parts:packet capture module,data preprocessing module,feature distribution module,sub-network detection module and decision fusion module.By applying the first two improved algorithms,intrusion detection can be carried out effectively in a friendly interface.In general,in order to solve the problem of conflict in multi-source data fusion,this paper provides a fusion algorithm from the data level,and improves the traditional D-S evidence theory to better apply to the fusion process.Finally,the two algorithms are applied to an intrusion detection system,and an intrusion detection system based on information fusion is implemented.