The Constructing For Knowledge Graph Of Micro-blog Personal Relationship

The microblog data contains a large number of personal relationships(for example,friend relationships,husband and wife relationships).However,the large amount of microblog information and redundant data make it difficult to obtain these complex personal relationships.In order to solve this problem,this paper proposes to Construct a micro-blog topic personal relationship knowledge graph in order to help users understand the complex personal relationship in the micro-blog environment.Firstly,the relevant person names are extracted from the microblog topic,and secondly the redundant information of these names in the microblog platform is used to extract the relationship of the characters,and finally the character relationship knowledge map under the microblog topic is constructed.The main research contents of this paper are as follows:(1)Relationship feature word extraction.In order to extract the relationship feature words as comprehensively as possible,this paper extracts the relationship feature words from three aspects:relation feature word extraction based on Dependency parsing analysis,relationship feature word extraction based on relative position information,and relationship feature words based on microblog topic Extract,then integrate and de-emphasize and eonstruct relationship feature group.(2)Optimization of Relationship Feature group and Construction of Personal Relationship Knowledge graph.In order to improve the effect of character extraction,this paper optimizes the relationship feature phrases.Firstly,a sliding window is established.The frequent feature set mining algorithm is used to extend the relationship feature words.Then,by moving the sliding window continuously,the total feature weights of the relationship feature group under the two sliding windows are compared to complete the optimization of the relationship feature word phrases.This paper research the method of constructing the relationship map of character relationship under the topic of microblog.Firstly,the name pair of the relationship to be extracted is obtained from the microblog topic,and then the information of the name pair is obtained through the microblog platform.The improved KNN method is used to extract the personal relationship triplet,and finally the Neo4j graph database is used to construct personal relationship knowledge graph.The experimental results show that the method of constructing the personal relationship knowledge graph of the microblog topic proposed in this paper can effectively represent the relationship of personal in the microblog topic.This paper innovatively uses the sliding window method to extend and optimize the relationship feature word group,which improves the effect of personal relationship extraction.By obtaining the names of people in the microblog topic and using the redundant information of these names in the microblog platform,the personal relationship knowledge graph of the topic is constructed.Figure[16]table[8]reference[59].