基于加权关联规则挖掘算法的电子商务商品推荐系统研究

摘要

To solve the direct commodity rapid and accurate matching problem between electronic shoppers and merchants, the e⁃commerce commodity recommendation system based on mining algorithm of weighted association rules is researched. Ai⁃ming at the insufficiency of the classic Apriori algorithm,a new weighted fuzzy association rules mining algorithm is put forward to ensure the downward closure of frequent item sets. The work flow of the recommendation system was tested through the struc⁃tural design of e⁃commerce recommendation system,data preprocessing module design and recommendation module design. The hit rate is selected as the evaluation standard of different recommendation models. The contrastive analysis for the practical col⁃lected data was conducted with the half⁃off cross test method. The experimental results show that the hit rate of Top⁃N products in association rule set is significantly higher than that of the interest recommendation method and best selling recommendation method.%为了解决电子购物者和商家直接的商品快速、准确匹配问题，进行基于加权关联规则挖掘算法的电子商务商品推荐系统研究。首先指出了经典Apriori算法的缺点和不足，并提出一种新的加权模糊关联挖掘模型算法，以保证频繁项集的向下封闭性；通过对电子商务推荐系统的结构化设计、数据预处理模块设计、推荐模块设计，完成了推荐系统的工作流程测试；最后选取命中率作为不同推荐模型的评价标准，通过五折交叉试验法对实际采集数据进行了对比分析，试验结果表明关联规则集的Top⁃N产品命中率要明显高于兴趣推荐和畅销推荐法。