支持用户属性特征联合聚类的协同过滤算法

User attribute-based collaborative filtering recommendation algorithm with co-clustering

个性推荐技术,可针对海量的数据实现快速准确的推荐。目前,协同过滤算法作为主流推荐技术,却存在着冷启动、数据稀疏和可扩展性等问题。虽然联合聚类协同过滤算法在解决可扩展性与数据稀疏方面有一定的效果,实时性高,但很难获得全局最优结果。因此,文中提出支持用户属性特征联合聚类的协同过滤算法以联合聚类为基础,融合了基于内容的推荐算法的优点,并经进一步改进以完成高准确率的推荐。实验结果表明,该算法实时响应快,一定程度上克服了冷启动和数据稀疏问题,且推荐质量较高。

A user attribute-based collaborative filtering recommendation algorithm with co-clustering personalized recommendation technology,is used to realize fast and accurate recommendation for vast amount of data. As a mainstream technology,the traditional collaborative filtering algorithm exists cold start,data sparseness and scalability issues. Although the co-clustering-based collaborative filtering recommendation algorithm has some effect on solving the scalability and data sparseness,and has higher real-time effect,it is difficult to obtain global optimal results. This paper proposes a collaborative filtering recommendation algorithm,which is based on co-clustering and also combines the advantages of contentbased recommendation algorithm,and makes further improvements in order to complete the recommend with higher accuracy rate. The experimental results show that the algorithm has fast real-time response,and to some extent overcomes the cold start and the data sparseness problem,also makes higher recommendation quality.