Mixed-mode Content-based And Collaborative Filtering Recommendation Techniques

With the popularization of Internet and the development of E-Commerce, the recommendation systems of E-Commerce are being paid more and more attention. To address this issue, recommendation systems were proposed to suggest products and to provide consumers with information to help them decide which products to purchase. Under the increasingly fierce competition, E-Commerce recommendation systems can enhance E-Commerce sales by converting browsers into buyers, increasing cross-sells and building loyalty to prevent users losing effectively.Presently the recommendation systems gradually become an important part in E-Commerce system, and more and more related papers appeared in many conferences and journals. Content-based Filtering and Collaborative Filtering are two of those successful application of recommendation technology. However with the wide practice of E-Commerce and growing systems, the recommendation systems encounter challenges.Aimed at these problems and challenges,, this paper explored and researched the E-Commerce recommendation systems based on content-based filtering and users-based collaborative filtering. It proposed a mixed recommendation based on content and users collaborative filtering recommendation algorithm. On the one hand This involved the advantages of content-based filtering which enabled all items be filtered according to the similarities, especially the items can be filtered and recommended to the users even without any users' evaluation. And the problems can be avoided at the beginning. On the other hand, it also involved the advantages of collaborative filtering which ensured the density of users' grading number matrix and the accuracy of the filtering when there is too many users and evaluation, collaborative filtering forecasting , So the combination of the two CAN improve the performance of system.The experimental result shows that the mixed recommendation algorithm is prior to that based on users collaborative filtering in MAE while when considered other performance indexes such as the precision, recall and F-measure, etc it is also prior to the recommendation algorithm based on content filtering.