Selection of Network Feature Attribute Based On Classification Discrimination And Correlation Analysis

Thereare many parameter attributes for anomaly detection of traditional network. It limits the detection efficiency. The selection of feature parameters is limited by the balance of samples. For these problems, this paper studies and designs a selection algorithm of network feature attributes based on classification discrimination and correlation analysis, defines the classification discrimination of feature, and a quantitative selection strategy is established. The test analysis is developed through the Moore data set from Cambridge university. The experimental results show that without reducing the detection accuracy, the algorithm proposed in this paper can improve the detection performance of abnormal samples, and effectively reduce the time of detection modeling.