Research On Knowledge Verification Method For Knowledge Map Construction

Since May,2012,search engine giant Google introduced "knowledge graph" in its search page for the first time,the construction of knowledge graph gradually caused wide attention of the academic circles and the industry,at the same time many researchers regard knowledge graph as a future search engine.With the Microsoft,Baidu,Sogou and other search engine companies in a short span of a year have announced their knowledge graph products,knowledge graph research has moved from theory to practical application.Knowledge graph construction is a dynamic process.It needs to add new knowledge and update dynamic knowledge in time to improve and supplement the existing knowledge graph.The new knowledge comes from the information extraction system,which extracts the knowledge for multiple data extraction.So it is necessary to validate the new knowledge,and select the new knowledge with high credibility as the candidate set to construct the high-quality knowledge graph.In this paper,we focus on the related issues of knowledge verification in the process of knowledge graph construction,which is divided into six parts.In the first part,this paper analyzes the current research situation of the research object at home and abroad,master the forefront of research.In the second part,this paper studies the theory and method of knowledge graph construction and knowledge validation model construction,which is the important theoretical foundation of knowledge verification model for knowledge graph construction.In the third part,this paper studies the entity resolution technology for multiple data sources,proposes the entity attribute distinguishing algorithm in the entity analysis process,uses the probabilistic soft logic model to realize the entity resolution technology,constructs the PSL model through three aspects of entity relations,entity attributes and ontology constraints And the reasoning mechanism of the PSL model realizes the decision-making process of entity resolution,and the experiment proves.In the fourth part,a knowledge verification model for knowledge graph construction is proposed,and the core part is discussed.By extracting the reliability of candidate knowledge sets from the information extraction system,classifying the knowledge of multiple data sources,and using the entity resolution technique and the ontology constraint technique to generate the probabilistic soft logic model with weight logic rules to achieve knowledge validation step by step.purpose.The fifth part,summarizes the research of the article and the prospect of the future work.The research value and commercial value of knowledge graph construction is enormous.This paper studies and analyzes the knowledge verification from the perspective of knowledge graph construction,and obtains very good experimentalresults.The research of this paper has important theoretical and practical significance to the construction of knowledge graph.Practical significance.