NADE

NADE数据集是由意大利乌迪内大学创建，专注于研究药物不良反应（ADE）提取模型在面对否定情况下的性能。该数据集包含2348条来自社交媒体的文本样本，其中包含大量标注的否定ADE实例。数据集的创建过程涉及从SMM4H数据集中筛选和标注相关样本，以及通过人工和自动方法生成否定样本。NADE数据集的应用领域主要集中在提高ADE提取模型在处理社交媒体文本时的准确性和鲁棒性，特别是在识别否定表达和避免误判方面。

The NADE dataset was developed by the University of Udine in Italy, focusing on evaluating the performance of adverse drug event (ADE) extraction models under negation scenarios. This dataset contains 2,348 social media-derived text samples, including a large number of annotated negated ADE instances. The construction of the NADE dataset involved screening and annotating relevant samples from the SMM4H dataset, as well as generating negated samples through both manual and automated approaches. The primary applications of the NADE dataset center on enhancing the accuracy and robustness of ADE extraction models when handling social media text, particularly in identifying negation expressions and avoiding misclassifications.