Few-Shot ASQP (FSQP)

该数据集名为FSQP，专为少样本角度情感四元预测构建，其特点在于类别更加丰富，数据更加平衡，有利于进行有效的少样本学习。数据集不仅包含了隐含的角度项和观点项，而且在标注者之间严格的四元匹配F1分数达到了78.63%，这表明在标注过程中有着高度的一致性和可信度。该数据集的规模包含12,551个句子和16,383个四元组，其任务定位于角度情感四元预测（Asqp）。

This dataset, named FSQP, is specifically constructed for few-shot aspect-based sentiment quadruple prediction. It is characterized by richer categories and more balanced data, which facilitates effective few-shot learning. The dataset not only contains implicit aspect terms and opinion terms, but also achieves an inter-annotator strict quadruple matching F1 score of 78.63%, which indicates high consistency and credibility during the annotation process. Comprising 12,551 sentences and 16,383 quadruples, this dataset targets the task of aspect-based sentiment quadruple prediction (Asqp).