Social And Comment Text CNN Model Based Automobile Recommendation

With the fast iteration and development of Web technology,on this basis,the development of online social networking has become a hot spot in people's lives.In the thirteen five period,the implementation of the national large data strategy and data resources to open and share the promotion of more and more Internet applications began to integrate with modern life,a large number of users through the Internet to provide a number of platforms to develop their own range of activities.With the rapid development of the city,work,self-drive tour,ferry the children to and from school and other activities increasingly frequent,the purchase of car demand gradually increased.However,this special commodity for the car,unlike the ordinary daily necessities,because it has its own high complexity,low frequency of transactions and consumption characteristics,so in the recommended task process,will encounter user-related areas of lack of knowledge,user history Low transaction data,high complexity of the product itself.In order to deal with the difficulties and challenges encountered in the recommended tasks,this paper is devoted to the design of the scientific and effective model by combining the social information and commentary on the social information on the purchaser's network and the characteristics of the comment data under different cars Recommended for cars.The paper starts from the above two aspects of information,in order to improve the recommendation accuracy as the goal of the individual car recommendation in-depth study,the main innovation and research results are as follows:(1)In order to better understand the user's rating behavior,need to tap the user's personalized features.The paper starts from two aspects,one is the social environment on the user's impact,in a specific model to build the needs of the purchase of the needs of the circle,in this circle based on the analysis of the user's personal preferences and preferences similarity two social factors,combined Scoring data and quantifying its social relations according to the existing calculation methods.Second,the impact of the text on the project,through the depth of learning technology to build a four-layer convolution neural network to learn the text features;(2)This paper puts forward a new automobile recommendation model SCTCMAR(Social and Comment Text CNN Model based Automobile Recommendation),which incorporates the social influence and comment text feature,that is,the personal preference and preference similarity of the user and the text feature of the project,And finally these three factors into the recommended model reflects the social factors and comments on the text of the characteristics of the decision-making process,improve the accuracy of the car recommended;(3)Using the matrix decomposition technique to train the latent feature of the user and the car,and then through the matrix merging technology to obtain the prediction preference matrix and finally to the target user to provide car recommendation.The results show that the proposed SCTCMAR vehicle recommendation model can carry out effective vehicle recommendation and have good automobile recommendation performance.(4)In order to verify the usability of the proposed algorithm,the SCTCMAR vehicle recommendation model was used to demonstrate the SCTCMAR model's efficient use in our automotive "Specialized Color Integrated Design System for the Automotive Industry".