Research On Expert Recommendation System For Security Domain Based On Machine Learning

As the country vigorously advocates and promotes innovation,encourages the spirit of innovation,and attaches great importance to the innovation and development of small and medium-sized enterprises,the technological innovation and reform of small and medium-sized enterprises are receiving more and more attention.As an important part of our domestic research and development cooperation,SMEs need to obtain support from the government and universities in terms of technology,capital,policies and talents in the reform and development.In fact,when SMEs encounter technological innovation and reform problems,they face tens of thousands of experts in related fields,and it is difficult to find suitable personnel for consultation in a short time.In response to this problem,based on data mining technology,this paper designs a security recommendation system for small and medium-sized enterprises.The system is mainly for the security field,including traffic safety,food safety,information security,production safety and so on.When the enterprises in the security field have technical requirements,they can match the similarity according to the requirements,find the most suitable and most suitable security experts in the enterprise,and save the cost of enterprise search,and it is more conducive to the wisdom of the enterprise and experts.Firstly,this paper studies the characteristics of the existing expert recommendation algorithms,and combines the characteristics of the security field,and puts forward the idea of using the patent information in the security field of the patent website as the latest research results of experts,and designs a web crawler that can capture the information of expert papers.The program is mainly based on the “Shanghai Science and Technology Innovation Resource Data Center” website,based on the Wanfang database to search for patent information with“security” as the key word.Then the patent information is preprocessed,and the word vector space model is constructed by using the Word2 Vec model.Then the convolutional neural network algorithm is used to classify the expert paper information.Secondly,in order to solve the cold start problem,enterprise users in the security field participate in the system design,and the enterprise users score the satisfaction according to the specific situation of the experts to judge whether the experts meet the needs of the enterprise and the industry.The system adjusts the recommendation results based on the company's rating.By optimizing theNon-negative matrix factorization algorithm of the recommendation system,the recommendation accuracy rate is improved,and the recommended content is displayed.Finally,based on the above research,the expert recommendation system in the security field is implemented,and the system is tested based on MoviesLens data.The test results show that the proposed machine recommendation system based on machine learning in this paper has certain feasibility and effectiveness.