Research On Semantic Recommendation System Derived From User Experience And Professional Evaluation

The development of mobile Internet enriches people’s production and life.The emer-gence of Internet services such as social networks,shared economy,mobile payment and e-commerce has brought convenience to people’s lives and made information overload more and more serious.The recommendation system is one of the effective ways to solve these problems.The recommendation system can model the users or products according to the user’s behavior,product attributes and other information,so as to recommend the valuable information and related products to the users.In recent years,tensor decom-position,text mining technology and cross-domain recommendation are widely used in the recommendation system,in order to alleviate or solve the sparsity problems in the traditional recommendation algorithm,improve the recommendation performance and recommendation quality.At the same time,with the improvement of quality of life,more and more users focus on the user experience of product and the authenticity and reliability of product information.Therefore,this paper mainly studies the method of extracting user experience information from comments and the cross-domain recommen-dation method using professional evaluat.ion information and user experience information.The main works of this paper are listed as follows:(1)This paper proposes a method of user-experience extraction for recommendation system.The formal description of user experience is given based on the user experience features.The detailed process of user experience extraction algorithm is given.The experimental results show that the proposed method has good accuracy and diversity at the same time.(2)This paper proposes a cross-domain recommendation algorithm based on tensor decomposition.After analyzing professional evaluation,a formal description of profes-sional evaluation is given,and a cross-domain joint recommendation model using user experience and professional evaluation is proposed.This model is an improved incremen-tal tensor decomposition model based on the tensor decomposition model.Experimental results show that the proposed model can effectively improve the accuracy and efficiency of the recommendation.(3)Designing and implementing a prototype system of recommended systems from user experience and professional evaluation.Test in real application scenarios,with high accuracy and good computing performance,the prototype system can be widely used in product recommendation with reviews and evaluation.This paper mainly studies the user experience mining method based on reviews and the cross-domain recommendation algorithm based on tensor decomposition.The experimental results show that the proposed algorithm can accurately and fully mine tlie product review information and improve the recommendation quality effectively.