Collaborative filtering of network malicious interference signals based on feature clustering
To address the network security issues caused by malicious interference signals,a collaborative filtering method for network malicious interference signals based on feature clustering is proposed.Based on the characteristics of network malicious inter-ference signals,the combination of envelope undulation,time-domain moment skewness,and time-domain moment kurtosis is used to describe the characteristics of network malicious interference signals.The FCM feature clustering algorithm is used to classify signal sample points with similar features into the same cluster according to the Euclidean distance from each network malicious interference signal feature point to the cluster center,completing the clustering of network malicious interference signal features;The feature clus-tering results are projected onto the feature subspace using the network signal matrix,and the principal component feature values are obtained using principal component analysis to divide the subspace into normal signal and malicious interference signal subspaces.The residual signal squared prediction error in the malicious interference signal subspace is compared with the filtering threshold of the malicious interference signal,and the malicious interference signal is determined and filtered.The experimental results show that this method can effectively filter out malicious network interference signals,obtain relatively pure network signals,and have high filtering accuracy,with fewer cases of missed and false positives.
feature clusteringmalicious interference signalscollaborative filteringFCM clustering algorithmfiltering thresh-oldresidual signal
万方数据
维普
