融入项目类别信息的协同过滤推荐算法

协同过滤技术在电子商务领域得到了广泛的研究和应用,但是随着互联网的迅速普及和电子商务网站规模的急剧增长,用户评分的极端稀疏性导致协同过滤方法的推荐质量不高.提出了一种融入项目类别信息的协同过滤推荐算法,结合项目的类别信息为活动用户筛选出候选近邻集合,在候选近邻集合内综合利用项目的评分信息和类别信息对未评分值进行预测,最后依据用户实际评分和预测评分计算出活动用户的最近邻集合并进行推荐.实验结果表明,该算法具有较好的推荐准确性和实时性.

An Algorithm of the Integration of Collaborative Filtering Recommendation With Item Category Information

Collaborative Filtering technology has a wide range of research and application in the field of E-commerce.But with the rapid popularization of the Internet and the sharp increase in the size of E-commerce sites,severe sparseness of user ratings make the quality of recommendation system decline.In this paper,an algorithm of the integration of collaborative filtering recommendation with item category information is proposed.And this method combine item category information to select the candidate neighbor se...