摘要
推荐系统是利用用户的历史偏好信息实现个性化服务的系统,它已经成为电子商务和信息获取领域中的重要应用。文中提出了一种通过模糊聚类的方法将项目属性特征的相似性与基于项目的协同过滤推荐技术相结合的推荐方法,首先应用模糊聚类技术对项目聚类,得到项目在属性特征上的相似关系群,然后与用户-项目评分矩阵中的协同相似关系群组合得到综合相似关系群,最后,利用综合相似关系群为目标用户推荐项目。实验结果表明,该方法不仅可有效改善基于项目的协同过滤推荐算法面临的“冷启动”问题,而且确实提高了推荐系统的推荐精度。
Recommender system uses customers' historical preferences to realize personal service. It has become an important application in E - commerce and Information access. This paper presents a new recommendation method, which combines the similar relation in attributes and characters of items with Item - Based Collaborative Filtering recommendation technology by fuzzy clustering method. Firstly it applies fuzzy clustering method to cluster the items and gets the similar group in the attributes and characters of items. Then this paper forms the synthetical similar group by combining the similar group with the collaborative similar group in user - item rating matrix. At last this new method makes the recommendation for active customers based on the synthetical similar group. The experimental results show that this method can not only efficently improve the Cold -Start problem of Item - based Collaborative Filtering recommendation algorithm, but also provide better recommendation results than that of Item - based Collaborative Filtering recommendation algorithm.
出处
《计算机仿真》
CSCD
2005年第8期144-147,共4页
Computer Simulation
关键词
电子商务
推荐系统
协同过滤推荐
模糊聚类
项目相似群
Electronic commerce
Recommender system
Collaborative filtering recommendation
Fuzzy clustering
Similar item group