A Intrusion Detection Algorithm Based on Improved Slime Mould Algorithm and Weighted Extreme Learning Machine

Aiming at the problem of intrusion detection, the extreme learning machine is not effective in processing imbalanced data. Using Levy flight to improve the slime mold optimization algorithm, an improved slime mold algorithm is proposed to optimize the weighted extreme learning machine intrusion detection algorithm. The improved algorithm can jump out of the local optimum with a higher probability, and the weighted extreme learning machine has the advantages of short training time and good generalization performance. It increases the weight of minority attacks and makes the detection rate of rare attacks in network attacks higher than that of traditional machines. The learning method has been greatly improved. Global optimization is performed on the input weights and biases of the hidden layer in the weighted extreme learning machine to avoid the algorithm from falling into the local optimal solution and realize the classification of the intrusion detection data set. Experiments show that the detection rate and classification accuracy of this algorithm for rare attacks are improved, and the false alarm rate is reduced.

I. Introduction

With the rapid development of Internet technology, traditional defense measures such as firewalls and encryption technologies can no longer resist the endless attacks from hackers. As a dynamic and active defense measure, intrusion detection can be more effective than traditional defense technologies. Good detection of whether the data flow in network transmission is abnormal is of great significance to ensure network security, so it has always been the focus of research in the field of network security.