基于用户模糊聚类的个性化推荐算法

Users' Fuzzy Clustering Individuation Recommendation Algorithm

针对目前协同过滤算法中存在的项目相似性计算不准确问题,提出运用模糊聚类技术对用户进行聚类,将单个用户对项目的评分转化为用户相似群体对项目的评分,构造密集的用户模糊簇-项目的评分矩阵,并结合项目自身的类别属性特征对相似性计算的影响最终完成项目相似性的计算,以提高项目相似性计算的准确度,实验结果表明,可提高协同过滤推荐算法的推荐精度。

This paper uses the fuzzy clustering algorithm in item's similar relation in order to resolve traditional method's disadvantage in this aspect,the method transforms the single users' preferences of item to similar user groups' preferences of item,which forms the dense preferences of user fuzzy-item,combines the similar relation in attributes and characters of item,it will be more correctly.Experimental result show that this method can provide better recommendation results than traditional collaborative filtering recommendation algorithm.