ICDM 2020 Knowledge Graph Contest Dataset

该数据集由行业解决方案专家精心挑选了500篇来自Instagram的最新文章组成，以确保文本在语言正式性、多样性以及实际应用知识深度方面均达到高标准。这些文章专注于消费者事件原因提取，数据集包含了用于训练的标注文章，并设有独立的在线测试集。该数据集主要关注五种事件类型：消费者关注、消费者兴趣、消费者需求、消费者购买以及消费者使用。任务是对这些消费者事件的原因进行提取。

This dataset comprises 500 recent Instagram posts curated by industry solution experts, ensuring the texts meet high standards in terms of linguistic formality, diversity, and depth of practical application knowledge. These posts focus on consumer event cause extraction. The dataset includes annotated posts for training purposes and a separate online test set. It primarily covers five types of consumer events: consumer concern, consumer interest, consumer need, consumer purchase, and consumer use. The core task of this dataset is to extract the causes of these consumer events.