Location-Based Personalized Recommendation System In M-Commerce

The development of wireless mobile communication technology and mobile ter-minals has brought about tremendous changes to people's lives. Much attention has been devoted to Mobile Commerce because it can effect transactions and provide ser-vices anytime, anywhere. Location Based Service is considered as a killer application in M-Commerce. However, users are still in the face of the problem, information overload, even in the LBS. At the same time, the limitations of users' time, energy, environment, screen of mobile terminals will make this problem worse.Recommendation technology can solve the information overload effectively and increase the viscosity and cross-selling ability of web sites, but the recommendation technology of traditional electronic commerce can not be transplanted to M-Commerce directly, because there are many characters in M-Commerce that differ from traditional electronic commerce. First, location relative, emergency and access any time anywhere are the unique value of M-Commerce; second, there are many differences in technology, service characteristics and business simulation for both of them. These differences determine that the recommendation algorithm of M-Commerce must fill the following needs: 1) Response quickly for the changes of user interest and can handle short-term interest. 2) Location-sensitive and can handle great changes of item properties. 3) No user ramp-up problem.From such requirement, recommendation is treated as the process of under-standing users'interest, narrowing optional space and helping users to make decision in this article. In the eyes of decision-making, recommendation is to select an item suited to the needs of user from limited optional ones, where each item is composed of several attributes. So the recommendation problem could be transformed to mul-ti-attribute decision-making problem. The representation model in this article is based on the case-based model, which has been modified to adapt the need of recommenda-tion in M-Commerce. On this basis, the method of TOPSIS is used to build this loca-tion-based personalized recommendation algorithm.In order to test the feasibility and effect of the model and algorithm, a testing system has been built which simulates recommending favorable hotel rooms for users in mobile environment. The result shows that the recommendation algorithm can fill the need of M-Commerce and has high recommendation precision.