A Collective User Preference Management System for U-Commerce

Abstract

In the area of electronic commerce, the personalized goods recommendation system is a very important research issue that raises user satisfaction, and increases loyalty towards the content provider. For this, the correct analysis of user preferences is essential, and most existing researches use a purchase history or a wish list. However, due to the rapid development of information technologies, commerce has progressed from e-commerce to U(Ubiquitous)-commerce. In the ubiquitous environment, computing devices of various types, including the mobile device itself, exist in user space; in addition, a broad range of information related to user preferences is generated while using these devices. Hence, if the information is efficiently managed, a more effective recommendation strategy will be established. This paper proposes a multi-agent based U-commerce system to efficiently collect and manage diverse context information that can occur in the ubiquitous environment. Therefore, a more personalized recommendation, which is reflected by various user preferences, is possible. A prototype was implemented in order to evaluate the proposed system, then, through results, the existing recommendation method is compared and the effectiveness of the system is confirmed.