Intrusion Detection Feature Selection Method Based On Improved Fireworks Algorithm

With the increasing popularity of the Internet,the types and quantity of network attacks are increasing rapidly.At this time,Intrusion Detection System(IDS)comes into being.The principle of this method is to match the characteristics of captured network data with those in the database.Therefore,it is very important to select the feature subset suitable for comparison,and feature selection is essentially a combinatorial optimization problem,and swarm intelligence algorithm is an effective tool to solve this kind of problem.Among the swarm intelligence algorithms,fireworks algorithm is more suitable to solve the problem of feature redundancy because of its advantages such as convenience,parallelism and dynamics.In this paper,a new Selectable Improved Fireworks Algorithm(FSMIFWA)is proposed based on the characteristics of Fireworks Algorithm and feature selection method.Firstly,an improved algorithm based on fireworks algorithm is proposed.Aiming at the problem of uneven distribution of initial fireworks,the algorithm proposed an initial fireworks dispersion technique,which made the initial location of fireworks cover the solution space uniformly.Aiming at the complex calculation process of the selection strategy,an iterative fireworks fitness value screening technique was proposed to reduce the time complexity and avoid the selection of inferior sparks with similar optimization ability,so as to effectively improve the ability and search efficiency of the fireworks algorithm to find the optimal solution.Secondly,the improved algorithm is mapped to the feature selection problem,and the feature selection method in this paper-FSMIFWA is proposed.This method provides the process of encoding and decoding,and makes it cover all feature subsets with low computational overhead,easy to understand and simple to implement.Finally,an intrusion detection model based on FSMIFWA is constructed,which not only retains the ability of rapid optimization of improved fireworks algorithm,but also reduces the feature dimension and the time wasted on redundant features in intrusion detection,thus ensuring the security of data in the network.In order to verify the effect of the feature selection method proposed in this paper in intrusion detection system,KDD Cup 99 and NSL-KDD data sets are used to carry out experimental tests.Experimental results show that compared with the typical intrusion detection model based on traditional fireworks algorithm and improved ant colony algorithm,the feature selection method proposed in this paper can effectively reduce the feature dimension,improve the detection speed and accuracy,better analyze the network data characteristics,and has a certain stability and availability.