基于灰色关联聚类的协同过滤推荐算法

针对协同过滤推荐系统具有数据的高稀疏,高维度,数据量大的特点,本文将灰色关联聚类与协同过虑推荐算法相结合,构建了灰色关联聚类的协同过滤推荐算法,将其应用到协同过滤推荐系统中,以解决数据具有高稀疏高维度的特性情况下的个性化推荐质量问题。首先,定义了推荐系统中的用户项目评分矩阵,用户灰色绝对关联度,用户灰色相似度,用户灰色关联聚类。然后,给出了灰色关联聚类的协同过滤推荐算法的计算方法和步骤,同时给出了评价推荐质量方法。最后,将本文算法与基于余弦,相关分析及修正的余弦等协同过滤推荐算法在大小不同的数据集下进行了实验,实验表明灰色关联聚类的协同过滤推荐算法相较于传统的协同过滤推荐方法具有推荐质量高,计算量小,对数据大小要求不高等优点,同时在推荐系统的冷启动,稳定性和计算效率方面也具有一定的优势。

Collaborative Filtering Recommendation Algorithm Based on Grey Incidence Clustering

In view of the collaborative filtering recommendation system having the characteristics of high data sparse, high dimension, mass data, combining grey incidence clustering with collaborative filtering recommendation algorithm, we have proposed collaborative filtering recommendation algorithm based on grey incidence clustering which can be applied to collaborative filtering recommendation system for resolving the problem of high sparse data, high-dimensional data, mass data and personalized recommendation quality. Firstly, we present the definitions of user item rating matrix in recommendation system, user grey absolute correlation, user grey similarity, user grey incidence clustering. Secondly, computational method and procedure of the collaborative filtering recommendation algorithm with grey incidence clustering, and its evaluation have been presented, and furthermore, the evaluation recommendation quality method is also been given. Lastly, our algorithm compares with the collaborative filtering recommendation algorithm, such as based on cosine, correlation analysis and modified cosine in different scale of data sets, experiment results indicate that our algorithm, compared with the traditional collaborative filtering recommendation method, has the advantages of the high recommended quality, small amount of calculation and the data size requirement being not high. In addition, our algorithm has some advantages in the cold start, stability and computational efficiency of the recommended system.