Design And Implementation Of Recommender System Of Knowledge Platform

Nowadays,the department of experiment is important for the development of industry and company.Using computer technology to assist industrial production has become popular in modern industrial production.Fields like aerospace,shipbuilding,ordnance,and mechanical equipment will carry out long-term design,simulation and test in industrial production.As a result,lots of test data will be accumulated.The test data management system(TDM)is developed in order to manage these unstructured experiment data.With the development of modern industry,the TDM system also needs to keep pace with the times.The development of computer technology,such as artificial intelligence,machine learning and other technologies,provides a theoretical basis for the development of the next generation of TDM systems.The advent of the information age and the rapid development of Internet related technologies have made Internet users enter the era of information overload.Because of information overload,it is difficult for users to quickly find information of theirs interest or valuable to themselves in the face of massive information.The recommender system has been created in this context,and now personalized recommendation technology has become one of the most important means to solve information overload.This paper has designed and implemented personalized recommender system for knowledge service platform of TDM system on the basis of TDM system.This paper first analyzes the actual needs and application scenarios of TDM system,and proposes a recommender system based on social tagging system.In order to solve the semantic fuzziness problem of label clustering,this paper establishes a label co-occurrence network and studies the overlapping community detection algorithm.LCLS,a new algorithm of overlapping community detection based on line graph and spectral clustering is proposed.The main contents are as follows:(1)By analyzing the practical application requirements of the TDM system and the knowledge platform and combining the requirements of the time development of the national planning of the "Intellectual Institutes" of our country,this paper puts forward a personalized knowledge recommender system based on the knowledge platform for the personnel of the scientific research institute.(2)Based on the analysis of the actual application scenarios,this paper builds a knowledge recommendation system based on tag clustering.In order to solve the polysemy problem in the label clustering,a label co occurrence network is established and the community discovery algorithm is used in complex network science to cluster the label.(3)In this paper,a new algorithm LCLS for overlapping community discovery based on line graph and spectral clustering is proposed in this paper,aiming at the problem that the community discovery algorithm ignores the strength of the node relationship.Besides,the cost of the algorithm is too large and the quality of the community is low.The LCLS algorithm improves the calculation method of similarity and evaluation function for the network graph with weight value,and uses line graph to cluster the edges of the network to find overlapping communities.In addition,the use of spectral clustering greatly reduces therunning cost of the LCLS algorithm,so that the LCLS algorithm can be applied to a large scale network.Experimental results on real data sets demonstrate the effectiveness and advantages of LCLS algorithm.(4)Implementing the personalized knowledge recommender system.In the establishment of the recommender model,the time factor is introduced to dynamically reflect the user's interest change,thus improving the quality of the recommendation,and this is verified by the experiment.Moreover,in order to improve the system performance and facilitate business modeling,this paper introduces graph database to replace the traditional relational database.In this paper,a new overlapping community detection algorithm LCLS is proposed,based on which a knowledge recommendation system based on label clustering is designed and implemented.The implementation of knowledge platform recommendation system and integration with other TDM systems is of great significance for the development of TDM system.Moreover,this is also a new direction for the development goals of smart institutes in China's scientific research institutes.The research work in this paper has certain value in theory and practice.