Study on Attribute Reduction Method of Network Intrusion Detection System Based on Granular Computing

摘要

Based on granular computing theory, according to the problem of intrusion detection classification performance reduced by redundant attribute in high dimensional network data, an attribute reduction method of network intrusion detection system based on granular computing is given, the redundant attribute is removed under the condition of keeping the information integrity of original attribute set to reduce the attribute dimension of data. The example analysis indicates that this method reduces the training and detection time, and improves the computing efficiency of system in order to reduce the data storage, it provides a new idea for processing massive large data.