Improved Moth-flame Optimization Algorithm For Network Intrusion Detection

With the widespread use of the Internet,the Internet has penetrated into all corners of our lives.With the popularization of the use of the Internet,the security of the Internet has attracted more and more attention.As one of the main security technologies,network intrusion detection technology emerges as the times require.Network intrusion detection technology is mainly based on the data collected before,to detect whether the access data is data attack data.In this regard,the main research of this paper is as follows:(1)In view of the problem that MFO algorithm converges too fast and easily falls into local optimum,Binary Moth-Flame Optimization Integrated with Particle Swarm Optimization(BPMFO)algorithm is proposed in this thesis.This algorithm introduces MFO helical flight formula,which has strong local search ability.It combines Particle Swarm Optimization(PSO)algorithm with speed updating method to make the individual population move along the direction of global and historical optimum,and increase the global convergence of the algorithm,so as to avoid falling into local optimum easily.First,the improved PMFO algorithm is tested by basic function,then the original algorithm is converted into binary algorithm and BPMFO algorithm by sigmoid function.The BPMFO algorithm is tested by UCI data set.The experiment proves that BPMFO algorithm and PMFO algorithm have practical value.(2)In view of the characteristics of large amount of data,high data dimension and low recognition rate in current network intrusion detection,this paper considers applying PMFO and BPMFO algorithms to classifier optimization and feature selection of network intrusion detection.In the aspect of classifier optimization,this paper considers applying the Moth-Flame Optimization integrated with Particle Swarm Optimization to the weight optimization of weighted K-nearest neighbor algorithm to improve the solving ability of the algorithm.The results show that the PMFO algorithm has a good effect in time and accuracy when applied to weighted K-nearest neighbor algorithm.In the aspect of feature selection,this paper considers the application of Binary Moth-Flame Optimization integrated with Particle Swarm Optimization to feature selection.The results show that BPMFO algorithm has obvious advantages in accuracy,efficiency,stability,convergence speed and comprehensive performance of jumping out of local optimum when applied to feature selection of network intrusion detection.(3)In this thesis,we consider combining the results of classification optimization of weighted k-nearest neighbor algorithm with those of feature selection,and apply them to network intrusion detection.The experiment is divided into four groups.The first group uses feature selection and weight optimization,the second group uses weighted k-nearest neighbor algorithm for weight optimization,the third group uses feature selection,and the fourth group does not use weighted k-nearest neighbor algorithm for weight optimization and feature selection.The experiment proves that the Improved Moth-Flame Optimization algorithm is applied to weighted K-nearest neighbor algorithm for weight optimization and network intrusion feature selection.Timing can significantly improve the recognition rate and operation efficiency of network intrusion detection.