电子商务推荐系统中推荐模型的研究

Study of Recommendation Models in E-Commerce Recommendation Systems

在基于Web使用挖掘的推荐系统中,高效地预测用户的浏览模式一直是研究的热点,但是,目前仅采用关联规则挖掘技术的Web推荐系统在预测用户未来浏览模式时很难取得令人满意的结果。提出三种推荐模型以提高预测精度、减少响应时间,实验表明,通过三种推荐模型的组合能够显著改进推荐的准确率、覆盖率和匹配率。

In the personalized recommendation systems,how to forecast users＇ browsing patterns effectively is always an important and hot area.However,the association rules mining do not perform well inpredicting future browsing patterns at present.In this paper,three recommendation models are proposed,which can enhance the precision of forecasting and reduce response time.The experiment results show that the combination of the three recommendation models could improve the precision rate,coverage rate and matching rate effectively.