Research On Construction Of Microblog Hot Search Knowledge Graph Based On Crowdsourcing

With the continuous development of Internet technology,Sina Weibo,a new type of social media,is loved by the majority of netizens and has become an indispensable part of daily life.Weibo generates a large amount of data every day,especially when there is a lot of information hidden in hot search topics?In order to understand the intricate internal relations between information from massive information,a knowledge graph came into being.The knowledge graph manages fragmented information knowledge in a semantic and structured manner,and describes in detail the concepts,entities and their relationships in Weibo hot searches.How to conduct a more in-depth study on Weibo hot search,extract key information from Weibo hot search content and comments,use the extracted knowledge and its connections to build a knowledge graph,and form a rich semantic knowledge base is constructed in this paper The main problems to be solved by searching the knowledge graph on Weibo.This article mainly builds a microblog hot search knowledge map based on crowdsourcing,uses crowdsourcing annotation technology to identify named entities and the relationships between entities,uses Neo4j graph database storage to build a knowledge map,and also proposes a sentiment analysis based on microblog hot search knowledge map.The feasibility of using knowledge graph to analyze public opinion was explored.The main work of this article is as follows:(1)In view of the short text of Weibo,irregular language,high noise,and more complex Chinese Weibo context,crowdsourcing tagging has the advantages of high efficiency and low cost.A method of entity recognition based on crowdsourcing tagging is proposed.Use a large number of crowdsourcers on the crowdsourcing platform to efficiently identify named entities.First,the capability of the crowdsourcing tagger is evaluated in the crowdsourcing process to determine the capability value of each tagger;then the maximum expected value algorithm is used to evaluate the capability value of the crowdsourced tagger obtained from the evaluation and the temporary label generated during the evaluation process.Analyze and learn to filter out the noise;finally,correct the result of the microblog crowdsourcing labeling based on the optimized crowdsourcing labeler capability value to determine the final labeling result.(2)The construction of the microblog hot search knowledge graph was realized and stored based on the graph database Neo4j.Firstly,the rellated concepts of the graphical database are explained,and the related advantages and uses of the Neo4j graph database are emphasized.Then the Neo4j graph data was used to complete the construction of the microblog hot search knowledge graph,and the feasibility of the knowledge graph construction process was verified.Finally,on the basis of constructing a good microblog hot search knowledge graph,the Cypher language is used to query the knowledge graph according to requirements.(3)A sentiment analysis is performed on the Weibo hot search knowledge graph review layer.In the comment layer of the well-searched knowledge map of Weibo,the convolutional neural network combined with the recurrent neural network method and the BERT-based method were used to perform sentiment analysis on the comment layer of the search knowledge map of Weibo,respectively.Methods for comparative analysis.Finally,public opinion analysis is performed on the basis of the knowledge map.Experiments show that the method proposed in this paper can effectively construct a microblog hot search knowledge map,and conduct public opinion analysis based on the knowledge map.Therefore,this article has certain practical significance for the development of knowledge map and public opinion analysis.Figure[35]table[11]reference[56]...