Ground-Truth Dataset for Class Cardinality Comparisons

该数据集包含90个类别，跨越6个领域，通过各类别的组合评估来研究类别基数。数据集不仅包含领域内的类别组合，还包含跨领域的类别组合，并由MTurk标注者在封闭式和开放式环境下提供注释。规模上，数据集包含了4005个类别对（共90个类别），其任务是对类别基数进行比较研究。This dataset includes 90 categories spanning 6 domains, and is developed to investigate category cardinality through combination evaluations of various categories. It encompasses both intra-domain and inter-domain category pairs, with annotations provided by MTurk annotators in both closed and open-ended settings. In terms of scale, the dataset consists of 4005 category pairs derived from the total 90 categories, and is designed for comparative studies on category cardinality.