A Novel Cyber-attack Detection Approach based on Kernel Extreme Learning Machine using FR-Conjugate Gradient
Jianlei Gao, Jun Li, Hao Jiang, Yaobing Li · 2020
Billions of devices have been developed with communication technology especially cyber technology, which provides great convenience for our lives. However, they also have produced some huge information security risks. The detection algorithm of traditional intrusion detection system (IDS) usually has a poor generalization and inefficient performance, that is to say it is not able to identify new attack types and deal with training data set with huge samples. In this article, a novel detection algorithm of IDS based on kernel extreme learning machine (KELM) using FR-conjugate gradient (FRCG-KELM) is proposed to overcome these problems. What's more, a standard NSL-KDD data set is used to evaluate its performance. Through comparing with KELM, the serial experiments verify that the proposed method can achieve a better performance including the higher detection accuracy, higher detection rate, higher generalization capability, lower false detection rate and greater computation ability.