Design And Implementation Of Scholar Label Discovery System Based On Knowledge Graph

Today,the number of scientific researchers and their achievements has been increasing,which has promoted the development of science and technology.However,due to the large number and scattered distribution of academic resources,it has brought great challenges to the analysis and retrieval of academic information.In this paper,a scholar tag discovery system based on knowledge graph is designed by combining scholar data,recommendation system and clustering algorithm from multiple data sources.The system clearly and intuitively displays the scholar’s personal attribute tag,research interest tag and social tag.For the design of the system,the main research contents of this paper are as follows:(1)Design and construct the knowledge graph of scholars.This is the basis and basis for subsequent label recommendation.Based on structural scholar data from multiple data sources,this paper defines the relationships and attributes between entities.Determine their collaborators through the research results of scholars,design a clustering algorithm based on collaborators for entity fusion,and realize the storage and visualization of scholars’ knowledge atlas with the help of Neo4j graphics database.(2)Design and implement interest tag recommendation algorithm.Due to the quality problems of existing data labels,a label preprocessing method is designed and a label set that meets the research task of this paper is constructed.On this basis,using the constructed scholar knowledge graph and the attention mechanism,a scholar interest tag recommendation algorithm is designed,which is compared with the classical BM25,CNN,LSTM algorithms on the manually annotated data set to verify the effectiveness of this algorithm.According to the needs of the actual project,this paper designs and implements a scholar tag discovery system,which provides scholar search and tag search functions.It displays various kinds of scholar tag information in detail on the scholar’s home page,and integrates this system into the Landinn system that has been put into operation online,thus verifying the availability of this system.