The Construction Method Of Knowledge Graph In Netnews

With the advent of Web2.0 era,the Internet has stepped into the era of information co-construction from the era of information sharing.The rapid development of online news has greatly changed the media and habits of people to obtain news information in their lives.In the era of big data,the efficient whole-network retrieval of the increasingly increasing news data on the Internet is often hindered by its massive,high complexity and unstructured characteristics.In order to improve the interactive experience between users and the Internet,achieve higher quality retrieval,and finally realize the purpose of predicting the information required by Internet users,knowledge map came into being.Based on the crawled network news data,this paper extracts the entities in a single topic news,adjusts the relationships among entities in the entity relationship network through algorithms,and constructs the knowledge map under each news.First of all,this paper simply deals with the obtained original news data to obtain the preliminary entities and related relationships.Then,based on the three-way closure principle of the initial relational network,the concept of "Stabilization" relationship between entities was established,and the ratio of "Stabilization" of each edge to "Stabilization" in the general diagram was calculated.Based on this,an index to evaluate the weight of the relationship between entities was established.Then the optimized Kruskal algorithm is used to obtain the Knowledge Graph of part of the topic news,which is referred to as the Basic Knowledge Graph(hereafter referred to as BKG)in this paper.For the same news event,BKG with different entity types can be constructed from the data captured in different media and at different time points.Since different news data report the same event,these BKG must contain highly similar entities.Then,a large number of entities with high similarity are used to fuse the obtained BKG map,so as to obtain a comprehensive knowledge map under a topic.Through the above scheme,the real news text data obtained can finally effectively form the knowledge map under the network news data.Finally,through the evaluation of experimental effect,the validity and reliability of the construction scheme adopted are guaranteed.Therefore,the content studied in this paper has certain practical value for the construction of knowledge map in the field of news..