客户导向目录分割问题的改进算法

Improved Algorithm of Customer Oriented Catalog Segmentation Problem

客户导向目录分割问题假设顾客至少对目录中一定数量的商品感兴趣,计算目录覆盖的顾客数量,据此评估目录分割结果.现有的分割算法为了保证目录尽可能多的覆盖顾客,而忽略了目录分割结果的效用.针对该问题,本文构建一种新的数据存储结构CFP-Tree用于存储顾客交易数据,并提出一种新的算法Effective-Cover解决目录分割问题.该算法使用树深度遍历法选择目录产品.实验结果表明,该算法能够获得更好的目录分割结果.

The customer oriented catalog segmentation problem assumes that one customer is interested in at least a certain number of items in the catalog, and then calculates the number of customers that are covered by the catalog. Hence, the result of catalog segmentation is assessed according to it. In order to ensure that the catalog covers customers as many as possible, the existing segmentation algorithm ignores the effect of the results of the catalog segmentation. Aiming at this problem, this paper constructs a new data storage structure CFP-Tree for storing customer transaction data, and presents a new algorithm Effective-Cover to solve the problem of catalog segmentation. The algorithm uses tree depth traversal method to select catalog products. The experimental results show that the algorithm can obtain better catalog segmentation results.