Design And Implementation Of Recommended System Based On Wireless News Client

With the development of Internet, new news is increasing all the time. Users are facing growing news and information. It takes a lot of time and effort to find users interesting news. In this context, personalized news recommendation system is coming for this problem. Personalized news recommendation system analyzing the user behavior by users use news client. With analyzing the user information the system help users to get the news what user interested in.Based on the above background and analysis of existing problems, this paper analyzed the related theory and technology of machine learning in the recommendation system. And then for the specific case of the dolphin news system, I analysis and design recommendation system basing the wireless news client. The main research content is the following five aspects.In the first, I analyzed the application of the data mining technology in the recommendation system, and referenced to the related theory and technology about recommendation system from various industries, and concluded that the functional requirement analysis of news recommendation systems.In the second, I described the overall design of dolphin news recommendation system, and analyzed the advantages and disadvantages of various recommendation methods. So the design from the simple recommendation system model to the composite recommendation system model included saying the character of a BOW model, named entity model and ESA model used for news.In the third, I design clustering model for inactive users. It help inactive users can use recommendation system to improving news CTR.In the fourth, I filter the use results by user feedback model. It can reduce the number of user dislike news and improve the accuracy of recommendation result.In the fifth, based on the wireless client, I implemented personalized recommendation system. At the same time I introduce the system storage. Make it work faster and provide a guarantee for expansion.The design and implementation of news recommendation system allow users access to get news. It also increase users click-through rate and improve news exhibition rate. In the era of rapid development of Internet, the system will have a wider development.