ACRec Dataset

该数据集将推荐书籍与用户偏好的情感认知描述相映射，其中包含了从书评中挖掘出的对细粒度情感状态的表述。该数据集包括情感（A）、情感认知（AC）和认知（C）三种陈述，总计包含12,522条情感陈述、1,252条情感认知陈述和45,935条认知陈述。数据规模涉及1,000名用户、249,326个时间步骤以及142,892本书籍。该数据集的任务是基于用户的情感偏好进行书籍推荐。

This dataset maps recommended books to the affective-cognitive descriptions of user preferences, and includes fine-grained affective state expressions extracted from book reviews. It encompasses three types of statements: affective (A), affective-cognitive (AC), and cognitive (C), with a total of 12,522 affective statements, 1,252 affective-cognitive statements, and 45,935 cognitive statements. In terms of data scale, the dataset covers 1,000 users, 249,326 time steps, and 142,892 books. The downstream task of this dataset is book recommendation based on users' affective preferences.