Application of a Distance-weighted KNN Algorithm Improved by Moth-Flame Optimization in Network Intrusion Detection
Hui Xu, Ce Fang, Qianqian Cao, Chaochuan Fu, Lingyu Yan, Siwei Wei · 2018
Since the Network Intrusion Detection System needs characteristics including high precision, strong stability and high efficiency, this paper proposes a Distance-weighted K-Nearest Neighbor (KNN) algorithm improved by Moth-Flame Optimization (MFO). The proposed algorithm first adds distance-weights to the original KNN algorithm to calculate Euclidean distance, so as to improve the classification accuracy of KNN. The proposed algorithm then utilizes the MFO algorithm, which has the characteristics of strong local search ability, fast convergence and short time. The experimental results show that, the proposed algorithm has characteristics of high classification precision and good stability, and the running time is nearly 2/3 shorter than other compared optimization algorithms in the experiments.