Research On The Method Of Mining Periodic Groups In Temporal Network

Periodic clique mining is a basic operation on the temporal graph,which is used to mine periodic cliques in the temporal graph.Efficient processing of periodic clique mining helps to understand the periodic group behavior in the temporal network and predict possible future events.Existing periodic clique mining methods need to perform periodic calculations on all vertices and edges in the temporal graph when pruning.Secondly,when performing graph transformation operations,the periodicity of vertices and edges needs to be recalculated again.Graph transformation may lead to transformation.The later graph scale is greater than or equal to the unconverted graph scale,and both stages have inefficiencies caused by repeated calculations.This article has conducted in-depth research on the above-mentioned problems,and the specific research content is as follows.Firstly,a solution method based on the time stamp sequence on the edge is proposed.The basic idea is to first enumerate and then verify.Based on this idea,an efficient algorithm EMP is designed.The algorithm first enumerates the maximal cliques that meet the requirements,and then performs periodic verification on the enumerated maximal cliques.The specific operation of verification is to extract the set of timestamps on each side of the extremely large group,and count the time points that appear in the set.If the number of occurrences at a certain time point is equal to the number of extracted sets,put it into a new set,and finally judge whether the sequence in the new set has periodicity.This algorithm can efficiently mine the number of periodic cliques in the temporal graph,but its enumeration efficiency is greatly affected by the scale of the original graph.Secondly,three efficient pruning strategies are proposed to reduce the original image.They are EMP-Flag Vex pruning strategy,EMP-Flag Edge pruning strategy,and EMP-Flag Edge+pruning strategy.EMP-Flag Vex is a pruning strategy based on the degree of the vertex,EMP-Flag Edge is a pruning strategy based on the length of the timestamp sequence on the edge,and EMP-Flag Edge+ is a pruning strategy based on the length of the periodic subsequence.These three strategies can effectively prune irrelevant vertices and edges,improve the efficiency of enumeration operations,and thus improve the overall efficiency of the algorithm.Finally,the existing algorithm and the EMP algorithm were compared and tested on 10 real data sets.The indicators to be compared include the running time,the influence of different parameters on the running time of the algorithm,the pruning efficiency of the pruning strategy,the running time of the pruning strategy,and the running time.The experimental results verify the efficiency of the algorithm in this paper.