FEPF: A knowledge Fusion and Evaluation Method based on Pagerank and Feature Selection

In recent years, with the development of various knowledge bases, the fusion of multi-source knowledge bases is a hot and difficult problem facing the field of knowledge fusion. Due to the large differences in knowledge base structure, the efficiency and accuracy of fusion are not high. Proposed Graph Structure Fusion, a totally new knowledge fusion method based on PR(PageRank) algorithm and feature selection. This method constructs a network graph for entity content. The PR value of each node is used to determine the closeness of the relationship with the target word, and the PR value is used to select Relevant entities, excluding irrelevant entities to improve computing efficiency, and then from the perspective of graph structure, fusion of multi-source knowledge base. Experiments show that the average precision of the algorithm is 92.8%.