Research On Data Query Optimization Technology Of Graph Database

Graph Database is a new database system,which bases on the idea and concept of graph theory and can deal with complex relational network efficiently.It is widely used in social network,real-time recommendation,credit information system,artificial intelligence and other fields.Graph data processing is an important research direction at home and abroad.However,due to the abundant application scenarios,the massive data,the complex data structure and other factors,which cause the efficiency of data query and the quality of data low in the actual application of graph database.Therefore,how to effectively improve the query processing ability of graph data,which is the key issue of graph database research.The main research core of graph data query optimization technology is how to prune graph data quickly and effectively,and the scale of data can be reduced by preprocessing and constructing index.Finally,the data can be searched quickly and accurately.The paper elaborates on two aspects:Firstly,the paper presents ahierarchical method based on maximum complete bipartite graph,which deeply study the existing enumeration algorithms of bipartite graphs.The algorithm has the following innovations: Firstly,it proposes a divide-and-conquer strategy based on maximal complete bipartite graph,which reduces the size of candiate sets by constantly updating the constraints of maximal complete bipartite graph.Secondly,according to the structure characteristics of bipartite graphs,the paper proposes two efficient pruning strategies and the initialization strategy,which can reduce the search space by compressingthe original graph.Secondly,the paper proposes a graph index technology based on feature nodes.According to the center point and path length of the two parameters,graph data is pruned quickly.The index has the following advantages: Firstly,the degree of graph nodes and the time of queries are used to measure the selection of feature nodes,instead of relying on expensive frequent subgraph mining algorithm,which reduces the time of index construction.Secondly,according to the feature nodes,density clustering is used to segment graph data and tree index is established by clustering results.It can reduce the size of querying data.Experiments show that the graph data query technology of the paper can improve the query efficiency of graph data,reduces the search space,improves the quality of service of data query and reduces the expenditure of index space.