Research And Implementation Of Graph Data Retrieval Technology For Multi-keywords Combination Query

With the popularization of knowledge graph and social networks,large-scale graph data has been widely used.The query of graph data has gradually drawn people's attention.Keyword search is one of the most basic search methods.This paper studies a multiple keywords combination of query graph data retrieval technology,by entering multiple query keywords to get the query results.There are many kinds of query results in the problem of keyword search.In this paper,the tree is used as the query result.The tree can link multiple query keywords.The root node of the tree can be connected to any node in the tree.So the path between the root node and the keyword node can be used to understand how the keywords are related.Mining the hidden links between the keywords helps to provide further search directions.The main work of this dissertation includes the following two parts.1)On the basis of BACKWARD algorithm,the same node may be traversed by the same query keyword multiple times.This traversal is not necessary.A pruning method is proposed to avoid such repeated traversal.A keyword search algorithm based on distance pruning KSDP is proposed.KSDP mainly improves the extension strategy part of BACKWARD algorithm.The experimental result shows that KSDP algorithm reduces the query time without changing the quality of the answer.2)On the basis of KSDP algorithm,this dissertation designs a keyword search algorithm based on distance index KSDI.KSDI algorithm uses graph partitioning and distance indexing.KSDI algorithm increases the granularity of traversed objects and takes advantage of the index.The experimental result shows that KSDI algorithm reduces the query time compared with the KSDP algorithm.