Efficient Computer Intrusion Detection Method based on Artificial Bee Colony Optimized Kernel Extreme Learning Machine

Abstract

With continuous development of computer networks, network attacks threat the information security of people’s daily life. For the protection against network intrusion behaviors, it is imperative to search efficient measurements to maintaining network security. Literature review indicates that taking the advantages of neural network, the network intrusion can be efficiently detected and the kernel extreme learning machine (KELM) can provide quick and accurate intrusion detection ability. The only parameter need be determined in KELM is the neuron number of hidden layer. Suitable neuron number will accelerate the training procedure. However, little work has been done to address the optimization of KELM. To address this issue, this paper proposed an effective method that uses the artificial bee colony (ABC) to optimize the KELM. With proper hidden layer neuron number, KELM could enhance the accuracy and speed of the intrusion detection. To verify the proposed method, experimental tests have been implemented in this work. The test result demonstrates that the proposed ABC-KELM can detect the network intrusion efficiently and its performance is superior to unoptimized KELM method.