结合用户交易情况的改进聚类算法

AN IMPROVED ALGORITHM FOR CLUSTERING ANALYSIS INTEGRATING WITH THE TRADE DATA

对通过URL-UserID关联矩阵得到页面聚类和用户聚类的算法进行了研究。指出了可以结合用户的交易结果来评价用户对商品页面的兴趣度,并给出了改进后的算法和计算过程,从而关联矩阵元素的权值能够更准确地反映用户对商品页面的感兴趣程度,使聚类分析结果更佳。

The page clustering and user clustering analysis based on the URL-UserID association matrix are studied. That the trade data of users can be used to estimate the interest of users in the merchandise page is pointed out. The improved algorithm and computation process are detailed. The right of element in the matrix can reflect the degree of users＇ interest in the merchandise page more accurately, which makes the result of clustering analysis better.