Modeling Of Network User Preference And Design Of Recommendation System

With the dramatic increase in the amount of network data, getting the required information from the mass of data has become an important technology. Especially in the areas of e-commerce, the correlation between the user and the goods has great commercial value. Recommendation system is a kind of filtration systems which is established to find such correlation. However, the key of ensures accuracy of recommendation is looking for similarities between users in the recommendation process. So a user preference modeling is particularly important. Researching modeling method of network user’s preference and recommendation method based on preference model can provide better information services to users with great significance.Microblogging is not only a Social networking site but also a media. It has been an important platform for online communication and information dissemination. The main work and innovation dissertation includes the following aspects:(1) This article briefly introduces the methods of analyzing network user’s preference and development and present condition of recommended techniques. Then the article clarifies the significance of this paper and illustrates several key technologies associated with this study.(2) This article puts forward two kinds of user preference modeling methods based on the user’s behavior information:the model based on fans relationship between users and the model based on user’s interactive information. Then designs two recommended options in accordance with the two models:the focus on recommendation system based on fans relationship between users and recommendation system based on user’s interactive information.(3) We select the microblogging data for the study, and design crawler system based on HttpClient and regular expressions. The crawler system achieves collecting microblogging data automatically.(4) In this paper, we use the data of Sina microblogging to experiment. We compare the recommended results through simulation and verify the feasibility of the recommended programs. We prove the feasibility of the model and recommendation system and provided a new idea of analyzing user’s preferences. The result indicates that this model can easily be used into the user behavior analysis in social networks. And this model can take advantage of relationship and interaction information between users. Finally, this article designs a microblogging recommendation system based on the recommendation program proposed before. This microblogging recommendation system is based on B/S development framework, and uses SpringMVC mode. This system also uses the MySQL database and uses AngularJS for front-end development.