基于云模型的时间修正协同过滤推荐算法

摘要

Aiming at the problem of data sparsity and time effects in the traditional collaborative filtering system,a Collaborative Filtering Recommendation Algorithm with Time-Adjusting Based on Cloud Model (CTCFR) is proposed. It creates the user's preference of items' attributes by using the cloud model, and adjusts the items rating similarity by establishing an exponential time function. Based on the data set from GroupLens project team,the experimental result shows that this algorithm can make the measurement of the items rating similarity more accurate and improve the quality of the recommendation better.%针对传统的协同过滤推荐系统存在的数据稀疏性和忽略时间影响的问题,本文提出了基于云模型的时间修正协同过滤推荐算法,利用云模型建立用户对项目特征属性的偏好度,并建立指数时间函数对项目的评分相似度沿时间维加以修正.算法采用美国GroupLens项目组提供的数据集进行实验.结果表明,该算法使得项目的评分相似度度量更趋准确,系统推荐质量有较明显的提高.