RARD

RARD数据集，由爱尔兰适应中心和德国莱布尼茨社会科学研究所等机构的研究人员创建，专注于研究论文推荐系统。该数据集包含57435086条推荐记录，记录了推荐算法的使用情况、特征类型、特征来源及推荐时间和点击情况。此外，还包含一个基于推荐点击日志的隐式项目-项目评分矩阵。RARD数据集使研究人员能够训练机器学习算法，进行离线评估，并研究Mr. DLib推荐系统的数据，而无需自行实现推荐系统。在科学推荐系统领域，该数据集具有独特性，提供了大量的隐式评分和多种推荐算法变体。数据集的应用领域包括推荐系统的算法评估、机器学习算法的训练以及用户行为的探索。

The RARD Dataset was developed by researchers from institutions including the Irish Adaptation Centre and the Leibniz Institute for the Social Sciences in Germany, with a focus on research paper recommendation systems. This dataset encompasses 57,435,086 recommendation records, which document the utilization of recommendation algorithms, feature types, feature sources, recommendation timestamps, and click behaviors. Furthermore, it includes an implicit item-item rating matrix derived from recommendation click logs. The RARD Dataset enables researchers to train machine learning algorithms, conduct offline evaluations, and analyze data from the Mr. DLib recommendation system without implementing a recommendation system from scratch. In the domain of scientific recommendation systems, this dataset holds unique value, as it provides a substantial volume of implicit ratings and diverse variants of recommendation algorithms. The application scenarios of this dataset cover algorithm evaluation for recommendation systems, training of machine learning algorithms, and exploration of user behaviors.