Application Research Of Improved Clique Search Algorithm In Finding Community

For the data of wireless city community discovery problem in group search(CS)algorithm process generation repetitive group,generate the redundancy,time algorithm complex degree higher,generating association between members of the group structure through higher and re defined group,and from the side of the optimized storage,pre edge processing,search to generate group three aspects,with special binary tree storage structure,the weight K selection sort,depth first traversal constructing algorithm T-CS.With special binary tree storage structure reduce the redundant edges and facilitate group structure search,with weight K selection sort effectively reduces the complexity of the sorting process time,depth first traversal of the search process is used to directly generate group greatly reduces the redundancy.For massive data overflow problem,combined with the MapReduce model,the MR-T-CS algorithm is proposed to realize the parallel data and weight sorting.Experimental results show that MR-T-CS algorithm not only can solve the running process a large number of duplicate clique problem,time cost is also greatly reduced,the massive data processing ability greatly enhanced the,clique generation representative greatly improved,in sparse graphs can be found in the CS algorithm found the subtle relationship,in the relationship between dense graph can be summarized CS algorithm for finding the relationship.