Research On Multi-keywords Search Based On Graph Structure

The development of information technology has dramatically changed the way people access resources.Among them,the keyword search on graph has attracted more and more attention.Multi-keywords search based on graph structure is to find all the subgraphs that can contain the query keywords on the data graph.Many documents are returned in the form of subtree,such as the minimum cost spanning tree.Recently,some researches also proposed algorithms for finding subgraphs.The heuristic algorithm is one of them,but its searching process is based on the weight of edges without prioritizing the similarities between nodes and keywords.Therefore,the existing algorithms not show the best connection between two nodes.At the same time,the existing methods also have some limitations on the ranking of top-k results.When calculating the edge weights,they simply average the weights of nodes at both ends without considering the influence of each query keyword.The sorting result is not satisfactory.In this paper,we propose a new multi-keywords retrieval method on the graph,which uses a new way to calculate the weights of nodes and designs an algorithm to search the sub-graph based on the content of the nodes.First,we get the set of nodes corresponding to each keyword,The content relevancy degree of nodes is chosen to join the result set in turn,and the influence of hub node is taken into consideration in the process of calculating the result edge set,overcoming the shortcomings of directly calculating the shortest path between two nodes.Eventually trying to return a more structured and non-redundant results.Secondly,this paper proposes a new calculation method to rank each result sub-graph,taking full account of the contribution of each keyword to the node content relevance,so as to return the optimal top-k results.Experiments show that the proposed method can quickly and accurately return the result to the user more in line with the user's query intent.