Association rule mining of network security monitoring data based on time series

Abstract

The traditional network security monitoring number association rule mining technology has low mining accuracy, so a time series based network security monitoring data association rule mining technology is designed. The preprocessing of time series to construct the corresponding time series frequency set, using SWFI - tree structure data storage model is set up, get after filtering and reorder the transaction data set, data sets of will be clean and remove invalid data and the remaining data formatting, finally USES the particle swarm optimization (pso) algorithm with limited data flow, recursive calculation of particle movement, build sparse list, complete monitoring data mining of association rules. The designed mining technology was used in the experiment with the traditional technology, and the experimental results showed that the designed mining technology was 23.22% more accurate than the traditional technology.