Personalized e-commerce system

Abstract

The phenomenon of information overloading is increasingly severe with the development of e-commerce websites. It is an urgent issue that how to make users find information they need efficiently in the huge information space and at the same time make e-commerce enterprises enhance their websites' attraction and sales effectively. The personalized e-commerce recommendation system is an effective method to solve the above problem. The coordinated filtering technology is one of the techniques that are most often used in recommendation systems. But traditional recommendation techniques still have many limitations in practice, such as data sparseness, cold start-up and scalability. In this paper, we prompt a new mixed recommendation model based on the analysis of the traditional coordinated filtering. We combine the item-based coordinated filtering with the item similarity analysis which based on items' attributes to find their near neighbors. Then we try to find the user's near neighbors in the items near neighbors. At last, the target user's interest is predicted according to his near neighbors' interest to the target item to get the top-K recommendation.