Research On Expert Recommendation Algorithm Oriented To Enterprise Demand

The expert recommendation algorithm for enterprise demands can recommend related experts for the enterprise based on the technical requirements of the enterprise.For the enterprise,when the company encounters technical problems in the development process,this recommendation algorithm can recommend the experts that meet the demands of the enterprise according to the technical requirements of the enterprise,saving the company's expert search costs and helping companies accurately dock expert wisdom.Firstly,this paper analyzes and summarizes the related literature.Based on the characteristics of the existing expert recommendation algorithm,this paper proposes a research idea of expert recommendation algorithm for enterprise demands.Demand is similar to the direction of expert research,but also consider the strength of experts'research ability.Secondly,this paper analyzes the characteristics of the company's demand data on the "Scientistin" website and integrates the company's demand data.A web crawler capable of capturing expert information was designed to mainly grab information published by experts in the CNKI database and information on research projects participated by experts.Thirdly,using the LDA model to extract the subject information of the company's demands and experts,and using the extracted subject terms and vector space model,respectively establish the expert feature model and the enterprise demand feature model.In order to improve the accuracy of constructing expert feature models,different weight values are set for the effective information of different parts of the paper.Experts are ranked according to the size of the similarity of the company's demand characteristics and expert characteristics,resulting in initial expert recommendation results.Fourth,in order to achieve the goal of not only considering the demands of enterprises and the research direction of experts,but also considering the strength of experts' research strengths,this paper further calculates the strength of experts' research ability based on the results of preliminary expert recommendations.The scientific research projects that the experts participate in can reflect the research direction of the experts and the scientific research practice of the experts to some extent.By calculating the similarities between the enterprise requirements and the scientific research projects,the similarities of the preliminary calculations can be further adjusted.Then,considering the publication time and the difficulty of publication,this paper constructs a model that can calculate the contribution value of expert papers,and adjusts the initial expert recommendation results according to the contribution value of the paper to produce the final expert recommendation results.Finally,this paper uses the actual enterprise demand data on the "Scientistin"website as an application scenario,and designs relevant examples.The algorithm studied in this paper is compared with the other two expert recommendation algorithms.The final recommendation results prove this Considering the consistency of the direction of the company's demand and the research direction of the experts,the expert recommendation algorithm that also considers the size of the scientific research strength of the experts is better.