Abstract: In this article, a multi-level association recommendation algorithm is proposed. Combined with product catalogs, the algorithm can automatically sel ect optimal matching granularity and make recommendation based on frequent items ets. The experimental results show that it greatly enhances online performance a nd quality. It can be successfully applied to personalization service in e-comme rce.

摘要： 提出了一种多层关联推荐算法，结合商品目录，自动快速选择最佳的匹配粒度， 进行基于频繁集的在线推荐。实验结果表明，算法大大提高了在线匹配的性能和质量，可成 功应用到电子商务的个性化服务中。

关键词: 推荐算法, 商品目录, Web挖掘

CLC Number: 

• TP399