Skip to Main Content
With the rapid development of E-commerce, people can get information easily from networks and customers have more choices, but at the same time it brings other problems. The vast amounts of information increase the burden for customers to purchase, they have to browse more unrelated information, and increase the time spent. To solve this problem and guide the customers' purchase in E-commerce, there needs to be an auto promotion system to help customers. In this research, we discuss the traditional collaborative filtering algorithm's, and propose a new item clustering-based collaborative filtering approach (ICSCFA). At first, the approach employs clustering items by support to decrease the nearest-neighbour space, and then gives the prediction of rate. The experiments have proven that the new approach increases the quality of clustering and is effective in relieving the extremely sparse customer rated matrix problem, enhancing the recommendation system's accuracy of prediction.