How to find appropriate automobile exhibition halls: Towards a personalized recommendation service for auto show

Abstract This paper proposes a novel recommendation methodology to guide visitors to find their proper automobile exhibition halls for auto show. In the proposed method, spatio-temporal features of visitors' behavior are first considered to construct their profiling, and then their interests are extracted based on visitors' clustering. Next, three modules including relevance module, quality module and integration module are developed for ranking visitors' preference of exhibition halls. Finally, highly desired exhibition halls are personalized and recommended to proper visitors. In the proposed modules, the relevance module is developed to measure the relationship of an automobile exhibition and a visitor, while the quality module is constructed to analyze the quality of each automobile exhibition. The integration module is to combine two modules above for recommending appropriate automobile exhibition. The proposed approach is well validated using a real world dataset, and compared with several baseline models. Our experimental results indicate that in terms of the well-known evaluation metrics, the proposed method can achieve more useful and feasible recommendation results, and our finding highlights that the proposed method can help both visitors to find a more appropriate automobile exhibition halls, and manage officers to reduce more management cost.