Building Events Knowledge Graph For Network Community

The emergence of the Internet has made it easier for people to access information.However,with the advent of the era of big data,how to obtain knowledge from large-scale heterogeneous information has become a difficult point.Since Google introduced the concept of knowledge Graph,it has become a trend to use knowledge-graph techniques to optimize the storage and representation of knowledge in a structured form.With the everexpanding scale of knowledge graph and applications,how to obtain deeper information of an entity and construct a knowledge graph with more solid stereoscopic images has become an important direction of current knowledge graph research.Taking the event type entity as an example,this paper proposes and constructs a network community knowledge graph that is event-centric and related to the expansion of event elements.In order to complete the construction of the knowledge graph,firstly,the data with high credibility such as encyclopedic entries and news reports are captured from the network community information,and then an event extraction model(LAL)based on LSTM and attention mechanism is established.Identify the events and related elements,and finally build an event-centric network community knowledge graph based on these event elements.In order to make the related knowledge involved in the knowledge graph more perfect,the similarity algorithm based on the maximum cosine and logistic regression of the word vector(Cosine-Logistic)is used to fuse the knowledge of other knowledge bases to supplement the basic attributes and relationships of the entity.Experiments show that the LAL and Cosine-Logistic models have achieved good results in both data sets and actual use.Using the data statistics and graphical query interface of Neo4 j database and CN-Dbpedia and other knowledge bases,the event knowledge graph supplements the knowledge loss of the general knowledge graph on event type entities,and can effectively record the development history of an event entity..This has great reference value for the specific understanding of the subsequent development and impact of an event and analysis event.