Network security analysis of weighted neural network with association rules mining

This article applies Co-S3OM semi-supervised learning algorithm to intrusion detection field and proposes specific semi-supervised network intrusion classification scheme. In accordance with different type of attack, different mark samples are selected as training set to complete initialization of three S3OM classifiers; marked sample data is expanded with coordinative vote by three classifiers. Test structure process is given in detail to use KDD Cup 99 data set to perform semi-supervised classification. It shows in test that intrusion classification model based on Co-S3OM is of high data sample marking rate and high intrusion classification rate.