The User Information-based Mobile Application Recommender System

Abstract

Smart phone and tablet PC users would like the variety of application contents in addition to web searching. Most of the mobile application stores tend to recommend other users applications merely because they have been chosen by many users. Thus, users who would like to choose applications that they personally prefer from hundreds of thousands of applications in mobile application stores need to spend a lot of time and efforts. This paper suggests a Collaborative Filtering(CF) method in utilization of users’ device ID information and the weight of certain demography information in order to minimize problems stated above and enhance the accuracy of recommendations. This method extracts preferred category information for each age bracket based on application categories and users’ age information, and decides the ranks in consideration of the similarity reflected in the weight of certain applications within the category based on the users’ information. The experiment shows that the proposed method improved the performance, which made accuracy of recommendation better than ranking-based method.