Research of E-commerce recommendation system based on multi-mode

Abstract

With the rapid development of e-commerce and the explosive growth of the information on Internet, the demands of recommendation system is being more and more urgent. In the past decades years, the research and application of recommendation system has been made considerable progress both in academia and industry. However, there are still a number of same problems that need to be improved constantly, such as sparse problems, cold-start problem etc. In this article, we design a multi-mode e-commerce recommendation system. By means of collecting a wide range of user's information, combined with a variety of recommendation algorithms which can learn from each other and finally comprehensively integrate all information into an overall result to provide users goods of service recommendations more accurately and comprehensively. © 2012 Springer-Verlag Berlin Heidelberg.