The Construction And Application Of Academic Knowledge Graph In The Field Of Information Systems

Academic papers coalesce scholars’ excellent research ideas.But under the current situation of massive paper data and information overload,it is a great challenge to fully explore the hidden knowledge in papers,add richer semantic information to the retrieval of papers,and help researchers get the required academic papers quickly and efficiently.Academic knowledge graph combined with recommendation technology can solve the above problems.With Google’s use of knowledge graphs for semantic retrieval,many excellent academic knowledge graphs have emerged,such as Ace KG.However,these large-scale knowledge graphs are mostly focused on the general and computer domains.The information systems domain and computer domain have crossover but different focus.Therefore,this paper focuses on the information systems domain,constructs academic knowledge graphs,and realizes the retrieval and related recommendation of papers based on these graphs.Firstly,we defined the data model of the academic knowledge graph of information system,and used the top-down approach to build the academic knowledge graph of information system,and added the semantic entity of "research method".A Text CNN model is trained to classify the titles and abstracts of papers as "research methods",and the effectiveness of the model is verified by comparing with SVM.Secondly,in terms of relationship complementation,the LDA topic model is used to represent the papers,and the similarity relationships of the papers are complemented by calculating the similarity degree with the defined meta-path rules;the Trans H model is used to represent the entities and relationships in the graph,and it works well in comparison with the Trans E model;the similarity relationships among authors are complemented based on the similarity relationships among papers.Finally,the upper layer application system is built based on the constructed knowledge graph.Based on Elastic Search,we realize the paper retrieval,display the relevant recommendations on the paper detail page based on the complemented paper similarity relations and author similarity relations,and visualize the knowledge graph of academic papers based on D3.js.In this paper,we study the construction and application of academic knowledge graphs in the field of information systems.A paper retrieval and recommendation system serving researchers in the information system field is built,and at the same time,it also provides a reference for the construction of knowledge graphs in other fields.