Research And Design Of Personalized Informatioin Service System Based On Context-Aware Recommendations

Along with the information technology and the development of the Internet, people gradually transit from the lack of information to the era of information overload, in order to solve the problem of information overload, the most representative of the solution is to classification direc-tories and search engines, but the two solutions are not yet fully able to satisfy people’s needs, and personalized information service system arises at the historic moment, this system through the establishment of user in-terest model for the user to provide personalized service, its core is to recommender engine. However the existing personalized information service system research does not consider the context information such as time, location, and companion. So this paper put forward a recommenda-tion algorithm using context information. The main things that were done are as follow.The first of all, analyze the shortness of user-based collaborative fil-tering algorithm and item-based collaborative filtering algorithm.Then propose the context-based recommendation technology based on the analysis of the disadvantages of the traditional collaborative filter-ing algorithm, giving the definition of context, expounding context access, context information modeling methods, finally dividing the context-based recommendation algorithm into three models:context pre filtering and after filtering, the context modeling, and then giving the corresponding algorithm respectively.And then the traditional recommendation algorithm and three con-text-based recommendation algorithm are designed and implemented us-ing the python language, and the accuracy, the recall rate, coverage rate, and popularity of recommend items of each algorithm are calculated, the three kinds of context-based recommendation algorithm and the tradi- tional recommendation algorithm in accuracy and ability of the long tail of mining are analyzed and compared.Finally designs and implements a context-based music recommenda-tion system, the architecture of context-based music recommendation system is designed, and analyzes its user-feature vector module, fea-ture-item module and filtering module. This paper expounds the way of how to development music recommendation system based on context us-ing the technology of wamp, and tests the context-based music recom-mendation system.