Research On Multidimensional User Model Based On Context And Personalized Recommendation Of Tourism Mobile Commerce

Tourism service is a kind of highly context-sensitive service. Active context should have influence on users’preference to some extent, when they adopt personalized tourism service recommendation. Main problems existing in recent research include lack of dimension weight and similar recommendation result. Some researchers extend user feature set with context data regardless of the influence of context on users’preference and recommendation results. Meanwhile, recent researches include only physical context, such as dimensions of time, location and temperature, which result in the similar recommendation to different users in the same context. The two questions above were studied in this thesis to improve personalized recommendation method, enhance personalized and adaptive degree of tourism mobile commerce, so users could take part in self-service through personalized recommendation system of tourism mobile commerce better.From the perspective of context, personalized recommendation of tourism mobile commerce will be studied with the example of tourist attraction recommendation. Based on the analysis of the research status of tourism mobile commerce and personalized recommendation, context theory and personalized recommendation method were used. Bayesian Network analysis was used to analyze users’ preference from the four dimensions of user profile, user historical behavior, active context and history context. A multidimensional user preference inferring model was proposed based on context modeling. Meanwhile, an experiment was conducted to test the recommendation quality of both the traditional and the novel method. It reveals that the personalized recommendation method based on multidimensional user model is better to some extent.