The purpose of this paper is study how to finding users with similar interests and recommending the tourist Spots to them effectively in big data, and use a framework for real-time recommendation. Due to the character of sparse data In field of tourism, the commonly method is according to user context, using a similarity algorithm clusters the users, Then suggesting the tourist spots to users with classical recommendation algorithm (such as collaborative filtering) in each cluster. But coupled with the rapidly increase data and more complex context, it did not work well as expect. This paper presents a novel algorithm framework which according to the real world characteristics of tourism data and based on the clustering and bipartite graphs theory. In this setting we aim to computing on-the-fly ranking of similarity tourist spots in multi-clustering, then use these result of each clustering similarity ranking to complete the real-time recommend precisely. Finally we show experimentally the accuracy of our approach with real-word data. The experiment results show that the proposed algorithm framework has a great improvement in both accuracy and real-time performance.
展开▼
