推荐系统由于其数据量庞大的原因,已经成为大数据领域研究的一个热点. 而协同过滤算法是推荐系统中最著名的算法之一. 传统协同过滤算法在利用评分矩阵进行推荐时,面临数据稀疏性问题,从而严重影响推荐的质量. 同时,推荐系统中存在大量的描述用户和产品属性特征的标签信息,把这些标签信息融入到传统的推荐算法中是解决稀疏性的一个有效方法. 因此,针对稀疏性问题,本文提出了一种结合标签和评分的协同过滤推荐算法. 该算法结合标签信息和评分数据共同计算用户之间或产品之间的相似性,进而为用户产生推荐. 实验结果表明,本文提出的算法可以有效解决数据稀疏性问题,同时可以提高推荐系统的准确性.The recommendation system has become the hot topic widely studied in the field of big data due to the massive amounts of data it contains. While the collaborative filtering algorithm is one of the most popular approach in the recommendation system. When making recommendations using the traditional collaborative filtering(CF)algorithms based on ratings matrix,we face the problem of sparsity that seriously impairs the quality of recommendation. Meanwhile,there is a large number of tags information that describe the attribute characteristics of users and items. Integrating these tags information into the traditional recommendation algorithms is a promising means to alleviate the sparsity problem. Therefore,to address the sparsity problem,this paper proposes a new collaborative filtering recommendation algorithm that integrates the tags and ratings,named UTR-CF. This algorithm utilizes the tags information and the ratings data simultaneously to compute the similarity between users or items,and then generate the recommendations. The experimental results indicate that the newly developed algorithm can alleviate the sparsity problem,and improve the accuracy of recommendation system simultaneously
