Design And Implementation Of Recommender System Based On The Mobile Internet

The current rapid development of Internet technology provides a majority of users with a lot of data which meet the user demand for information. However, a variety of overwhelming information lead to a fact that more and more users feel difficult to search their interesting information, with the phenomenon of so-called "information overload". The personalized recommendation technology is an effective method to solve this problem. It recommends possible information which users might be interested in, by analyzing historical and behavior information about the user to establish the interesting model. Currently, hybrid recommendation technology is widely used, it integrated all different kinds of factors with the algorithm which could be called depth customization. In this paper, we made some elaborate analysises about the hybrid recommendation algorithm, and built a recommendation system of news which based on that algorithm.The main works of the paper as follows:(1)At first, we proposed a method of recommend system which based on the mobile Internet, built user interesting model and news context model. And then we generated the semantic text by using LDA topic model training method. The system builds data dictionary semantic corresponence between the text and score, and recommends the keywords. In the end, the news will match the original library.(2)In the process of recommendation, according to the problem of sparse matrix and cold start, we clustered the dataset and fill it to achieve the line of start. By integrated the Ebbinghaus law into the matrix, we add factors of time to increase the quality of recommendation system, and improve the issue of user interest migration.(3)In order to solve the problem when the system meet the big data. We use a distributed framework. By using non-relational data HBase and relational database MySQL combination of data storage to meet the algorithms and functions to achieve the practical requirements of combining, through Ajax asynchronous data exchange technology, PhoneGap client cross-platform mobile development technology, Spring MVC framework, we designed a system solution based on the mobile Internet.In the end, we achieved the system solution, made the algorithm module and function module cooperated accurately, finished the development of client APP and news recommendation system.