Research And Application Of Improved Artificial Immune Network Algorithm In Network Attack Analysis

As an auxiliary means of network security defense,network attack analysis can analyze the network attack situation,so as to take effective defense measures in time.Artificial immune network has the characteristics of clustering and can accurately classify abnormal traffic,but there are problems such as low training efficiency and data set noise,which will lead to the deviation of subnetwork coverage.Therefore,this thesis proposes an artificial immune network algorithm based on cell division characteristics and applies it to the analysis of network attacks.The detailed work in this thesis is as follows.(1)By analyzing the research status of artificial immune network algorithm at home and abroad,this thesis considers that there are two defects in artificial immune network algorithm.Firstly,the training efficiency of the algorithm is low,and part of the data in the training process cannot be fully used due to the process of artificial immune network.Second,the noise of the data set will generate more invalid B cells when stimulating the sub-network during the algorithm training process,resulting in a deviation in the coverage of the sub-network and affecting the detection rate of the algorithm.Aiming at these two defects,this thesis proposes an improved idea using the characteristics of cell division,and demonstrates the rationality of the improved idea.In the process of demonstrating the rationality of the improved idea,the combination of the characteristics of cell division and artificial immune network is discussed.(2)In this thesis,by referring to the mechanism of cell division,according to the phenomenon that the characteristics of cell division are not consistent in the normal state and the aging state,the training process is redesigned,and an improved artificial immune network algorithm is proposed.In this algorithm,the concept of cell age is introduced,and the criterion for selecting data in the training space is expanded to the double combination of cell age and affinity.At the same time,cell age was used to simulate cell aging,and different methods were used to process data in different states during training.The improved artificial immune network algorithm proposed in this thesis optimizes the training process of the original artificial immune network algorithm,improves the training efficiency,and makes the generated network structure clearer.Finally,the CIC-IDS 2018 data set and KDDCUP 99 data set are selected as the experimental verification data set.The experimental data show that the proposed algorithm has strong robustness and high clustering accuracy.(3)In this thesis,the artificial immune network algorithm based on cell division is used to construct a network attack detection model,and the model is applied to the network attack analysis system.OPCUA and CIC-DDo S2019 datasets are selected for experiments to verify the overall performance of the model,and the implemented network attack analysis system is simulated to show the analysis ability of the system for network attacks.