论文部分内容阅读
随着 Internet 的普及和电子商务的盛行,智能推荐系统也应运而生.协同推荐是目前公认为最好的一种推荐技术,但其存在着一些不足之处,如:稀疏性、可扩展性和冷启动问题.本文提出一种混合推荐技术来克服协同过滤的不足.首先,通过引入多个数据源对评价矩阵进行平滑填充来解决数据的稀疏性问题.其次,采用从用户和项目两方面进行联合聚类来提高系统的可扩展性和精度.实验结果证明,该方法在很大程度上较传统的协同过滤方法推荐精度高,且在线推荐的速度快.
With the popularization of Internet and the prevalence of e-commerce, intelligent recommendation system emerges at the historic moment.Co-recommendation is one of the best recommendation techniques currently recognized, but there are some shortcomings such as sparsity, scalability And cold start problem.This paper presents a hybrid recommendation technology to overcome the deficiencies of collaborative filtering.Firstly, through the introduction of multiple data sources to smooth the evaluation matrix to solve the problem of data sparsity.Secondly, using two aspects from the user and project The joint clustering is used to improve the scalability and accuracy of the system.Experimental results show that the proposed method is more accurate than the traditional collaborative filtering methods and the online recommendation is faster.