Mobile Service Recommendation Based On User Consumption Behavior

With the popularization of mobile terminals and mobile communication technology,business homogenization has become an increasingly prominent problem forcommunication service providers. And the demands and interests can vary widely amongdifferent types of mobile users. The blindness in the initial choosing of tariff packagesleads to unsteadiness and shorter life cycle of customers. Moreover, these customerscontribute relative low ARPU values. Therefore how to provide targeted, personalized anddifferentiated telecom services to meet the specific demands of different customers hasbecome an urgent problem for telecom operators.On the basis of extensively study of various types of personalized recommendationsystem and user behavior modeling technology, this paper proposed a hybridrecommendation system to provide personalized telecom services recommendation service.The system solves the problems of data sacristy and scalability. To achieve this goal, thepaper proposes an user consumption behavior analysis method for the construction ofpersonalized ontological user profile. The profile was applied in the recommendationprototype system.On the basis of the existing consumption behavior analysis method, the paperproposed a user modeling algorithm. The algorithm can construct a personalized userprofile reflecting the user’s preference of tariff packages. To identify the priority ofrecommended items, the paper proposed a rule confidence calculation method based onuser preference model to identify the priority of recommended items. In order to build auser-based collaborative recommendation system, the paper proposed a user similaritycalculation method to get user groups of similar interest based on user preference model,in order to build a user-based collaborative recommendation system. The paper combinedthe rule-based and user-based collaborative recommendation system to providepersonalized telecom services in the form of web pages and external interface.