Research On Dynamic Graph Matching Algorithm Based On Increment

With the arrival of the big data era,a large amount of data stored in graph structure has been accumulated in the field of network analysis,and it contains rich information of practical value.Therefore,graph matching technology of analyzing and mining useful information emerges as the times require.When the data scale is larger and larger,the dynamic graph matching algorithm based on snapshot mode has a large number of redundant calculations when the data is updated frequently,which leads to the complexity of the algorithm matching time.In view of the above problems,the following research has been carried out:Firstly,considering the priority of multiple query subgraphs after the large-scale query map,adopt the core-forest-leaf decomposition strategy according to the node size of the query map,determine the priority order of query subgraph,and give the query graph nodes by path search.Secondly,in view of the problem of high memory consumption caused by the large data map scale,by generating the query map width priority search tree,the data map nodes are stored in the query node candidate set as embedded according to the top-down matching strategy,so as to build a compressed path index model for the query map.We also define the two types of index structures according to the query graph non-tree edge for the query node.Thirdly,for the problem of high time complexity of the dynamic graph matching algorithm on the large-scale data graph,we use the type of index structure of the query node satisfied by the added node and its neighbor node to delete the index without updating.Match within the area of the node index updated,update the previous matching result set,finally get the new matching result set and the updated query node index,and be used for subsequent incremental matching.Finally,the time efficiency of the above algorithm is verified and analyzed on different scales of data sets.