关键词: 个性化推荐, 顾客行为, 协同过滤

Abstract: The E-commerce leads to the moreintensive competition among the entities, and personalization recommendation would be the breakthrough to increase the customer's loyalty and promote the profits. The dynamic mining algorithm was proposed and improved to realize the personalization recommendation based on customers purchase sequences. The definition of time contraint was put forward considering the impact of the length of the times compartmentalization and the selflearning function of the system was realized. The result of an imitating experiment indicates that the recommendation method proposed is feasible and effctive.

Key words: personalization recommendation, customer’s behavior, collaborative filtering