PatentAP

该数据集名为PatentAP，用于通过分析专利声明及其依赖关系来预测专利批准。数据集包含了诸如专利类别、引用数量和新颖性得分等特征。在评估模型性能时，采用了曲线下面积（AUC）和宏观F1分数作为主要评价指标。该任务的目标是进行专利批准预测。

This dataset, named PatentAP, is designed for patent approval prediction by analyzing patent claims and their dependency relationships. It includes features such as patent category, citation count, and novelty score. When evaluating model performance, Area Under the Curve (AUC) and macro-F1 score are adopted as the primary evaluation metrics. The objective of this task is patent approval prediction.