- 期刊
A PSO-based Wavelet-core ELM for Abnormal Flow Detection
摘要
Abnormal flow detection is an effective approach to discover the covert data during the transmission process of mass data. However, there exist some issues to tackle such as the high complexity of network traffic data, Low detection efficiency and low accuracy. To solve these problems, we proposes an improved wavelet-core extreme learning machine based on particle swarm optimization. First, the particle swarm optimization algorithm is applied to determine the input weights and bias thresholds of the extreme learning machine, which effectively reduces the number of hidden layer nodes. Furthermore, wavelet kernel function is proposed to be the kernel function of kernel extreme learning machine. Then the topology of the KELM can be established, and can be applied to classify the abnormal traffic. We introduce overall-accuracy and F-measure for performance measure in abnormal flow detection. To verify the effectiveness of our work, we compare the approach with the representative algorithms, and experimental results show that the improved wavelet-core extreme learning machine based on particle swarm optimization has better detection performance.