The Submodule Of Theme Circle Found In EgoNet Of Expert Discovery Platform

The Institute of Political Science and Technology is a new economic form in knowledge society environment.The biggest difficulty faced by the government and enterprises in the promotion of Industry-University-Institute is the talent introduction.In academia,research literature is the result of academic activities with multiple experts completing cross-domain cooperation to achieve the same goal.Therefore,in the coauthor social network composed by co-author relationship,the experts of rich achievements will form a self-centered micro network—EgoNet,due to the wide range of research and frequent activities.Other experts in EgoNet have been divided into corresponding community circles according to their respective research fields.Therefore,more experts in different disciplines can be discovered by researching prolific experts' EgoNet,which greatly satisfying the need of experts of specific research fields for enterprises.As an important aid to the expert discovery platform,the expert EgoNet and theme circle discovery module will complete the EgoNet construction of prolific experts,as well as the discovery of the theme community circle in the coauthor social network.In order to realize the preliminary study of expert EgoNet and the discovery of community structure in the scientific co-authored complex network,taking the abstract content of expert literature as textual corpus data,the main research work is as follows: According to the network topology,the important nodes in the coauthor social network are marked with the compactness centrality as the measurement standard;Author-Topic Model was used to cluster expert literature abstract text,to predict the probability distribution of expert objects in the research field as the Topic,and construct EgoNet for important nodes by combining network topology characteristics;In the social network of scientific research co-authors,the improved Infomap algorithm model was used to finish community discovery,and according to the clustering results,themes were marked for the community structure to form a theme community circle,so that expert nodes were assigned to the corresponding theme circle,and the distribution results were presented in EgoNet.There will finally completed the design and implementation of the expert EgoNet micro-website system,which can be used to analyze the co-authors' scientific research and their main research directions.In egonet,the link strength between nodes in expert EgoNet reflects the similarity of experts in coauthors and research fields,and the distribution of expert nodes in the topic circle reflects the main research direction of experts,and also clearly explains the different technical support obtained by experts as the central node.The machine learning method – AuthorTopic clustering model is applied to the study of social networks,to avoid the only according to the network topology analysis,the physical structure of the network makes the expert object as a network node,to consider the realistic meaning of expert of social behavior for the system to provide services for experts have discovered,is more realistic.