Intrusion detection method based on sparrow search optimization deep extreme learning machine

Deep Extreme Learning Machine (DELM) has been successfully applied in many fields due to its good performance and strong generalization ability. In view of the low detection efficiency of the existing intrusion detection technology, DELM is introduced into the network intrusion detection. And in view of the large randomness of its initial parameters, an intrusion detection model RSSA-DELM based on reformative sparrow search algorithm (RSSA) optimized DELM is proposed.First of all, in the sparrow search algorithm (SSA), the position update formula of the sparrow finder and the sparrow guard is improved, which effectively avoids the SSA algorithm from falling into the local optimal and introduces a random walk strategy to disturb the optimal solution of the sparrow to further improve the sparrow search ability and increase the diversity of the population. Compared with the standard sparrow search algorithm (SSA), particle swarm optimization algorithm (PSO) and whale optimization algorithm (WOA) on the four test functions, the reformative sparrow search algorithm (RSSA) has faster convergence speed, higher convergence accuracy and good performance. Then use the reformative sparrow search algorithm to jointly optimize the weight and bias of DELM, and finally use the optimized DELM algorithm to classify and detect the NSL-KDD network data set. The experimental results show that RSSA-DELM has a higher detection rate than DELM, SSA-DELM, RNN and other algorithms, and the classification performance is improved by an average of 18%.