Research And Realization Of Mobile Terminals Personalized Application Services Push System

With the rapid increase of mobile terminals popularizing rate and the rapiddevelopment of mobile Internet and mobile communication technology, mobileapplications market in recent years, the number of applications in the mobileapplications market appeared explosive growth in recent years. Faced with a largenumber of mobile applications, users often need to spend a lot of time to choose theirfavorite application. To solve this problem, the mobile personalization applicationspush services have emerged.The personalized application services push system mentioned in this paper canautomatically push mobile applications to users based on their preferences, whichsaves the users’ time. The system can help operators to provide customers with betterservice and can promote the development of value-added services which bringsgreater benefits for operators.The main research works in this paper are as follows:First, this paper introduces the concept and composition of personalizedrecommendation and related theoretical knowledge. Also it researches therecommended techniques which often used in traditional personalizedrecommendation system, and analyzes the advantages and disadvantages of differentrecommended techniques.Secondly, it focuses on the analysis of collaborative filtering algorithm which isvery popular in recent years. And it analyzes the basic principles, the classificationand the advantages and disadvantages of the algorithm.Thirdly, it presents an improved weighted Slope One recommendation algorithm.Slope One algorithm is a simple and efficient item-based collaborative filteringalgorithm which was proposed in2005, however the algorithm still exists the problem of cold start. For the “new users”, the improved algorithm calculates the similaritybetween users by users’ personal information, then generates the users’ nearestneighbor by the similarity, and at last carries out recommendation through the score ofnearest neighbor users. For the “new items”, the improved algorithm calculates thesimilarity between items by their own property, then predicts score accordingsimilarity, and then carries out recommendation.Finally, the paper designs and implements the mobile terminal personalizedapplication services push system. The system consists of two parts, server and client.The part of server includes the personalized recommendation module and systemmanagement module. And personalized applications recommended modulerecommends applications to users by running the improved weighted Slope Onerecommendation algorithm. The system management module implements thefunctions of application management and statistics. As an interface for users, the partof client is mainly responsible for the application display. What’s more, there is abackground thread running in client, which is used to gather application usageinformation, and this information is the data source of server to recommendapplications.