Design And Implementation Of Scholar Search And Scholar Community Discovery System

Scholar search and scholar community discovery system is a retrieval tool in industry university research service system,which provides data support service and information retrieval function for industry university research service.Scholars in Colleges and universities undertake scientific research projects and apply high-value scientific and technological achievements to production,which can promote industrial upgrading and enhance competitiveness.But in the actual process of industry university research cooperation,there are some problems,such as the disconnection of academic and technological,the low conversion rate of achievements and so on.The existing retrieval system focuses on academic retrieval,and the retrieval object is single and the function is simple,so it is difficult to be applied in the actual industry university research project.Therefore,it has important application value to build a system that can quickly and accurately retrieve the cooperative objects and evaluate their capabilities.This thesis designs and implements the scholar search and scholar community discovery system,realizes the scholar search function and scholar community discovery function,provides the government and the enterprise with the decision support tool of industry university research,and solves the problem of talent introduction and achievement transformation.Scholar search can recommend scientific research talents for industry university research service,while scholar community discovery can dig out relevant research teams and provide more choices for industry university research service.In the process of system implementation,how to accurately represent scholars,effectively retrieve scholars and discover scholar community has become a technical problem.There are a lot of information implied in the cooperation between scholars,and the text-based scholar representation method has some limitations.Traditional community discovery methods focus on the structure of scholar network and ignore attribute information.In order to solve the above problems,the main contents of this thesis include:(1)This thesis proposes a scholar retrieval method based on word embedding.Word embedding model can learn semantic information,add network topology attributes to the model,and improve the accuracy of retrieval model.The process of scholar retrieval includes scholar node representation and scholar vector retrieval.Firstly,we construct the scholar cooperation network,and add the network topology information into the word embedding model to get the scholar node representation vector,which is composed of text vector and node vector.The text vector is generated by the text of the current node and adjacent nodes,while the node vector is generated by the adjacent nodes.Then the product quantization model is used to retrieve the scholar vector and return the relevant scholars.Compared with the traditional model,the retrieval model based on word embedding can improve the recall while ensuring the precision.(2)A topic based local scholar community discovery method is proposed.Scholars can contain multiple research topics and form different research circles.The traditional community division method is often inaccurate.In this thesis,the subject information is added to the scholar cooperation network.According to the user input,the system can return the scholar community with different topics.The realization of scholar community discovery can be divided into three steps: first,expand the scholar node into scholar community according to the retrieval subject;secondly,prune the community with k-truss,delete the discrete nodes,and find the scholar community with close connection;finally,rank the scholar community according to the innovation ability evaluation of the scholar community.After the annotation data test,the community discovery model implemented in this thesis can achieve good results.(3)The system of scholar search and scholar community discovery is designed and implemented.The system module can be divided into background management module,user retrieval module and data storage module.The system mainly realizes the functions of scholar retrieval,scholar community discovery and information display.The staff can input the retrieval subject in the system,and the system outputs the relevant scholars and scholar communities,and display the relevant information.Finally,the algorithm test,function test and performance test are carried out.The test results show that each functional module of the system can work normally,and the online retrieval performance can meet the expected requirements.