E-commerce QA Dataset
收藏arXiv2025-09-30 收录
下载链接：
https://github.com/gsh199449/productqa
下载链接
链接失效反馈官方服务：
资源简介：
该数据集是从一个真实世界的电子商务网站收集的大规模数据集，包含了问答对、评论以及产品属性信息。平均而言，问题包含9.03个单词，而答案则包含10.3个单词。在训练样本中，有78.74%的样本拥有超过10条相关评论。该数据集涵盖了469,953个产品和38个产品类别，旨在支持产品感知的答案生成任务。
This is a large-scale dataset collected from a real-world e-commerce website, which includes question-answer pairs, customer reviews, and product attribute information. On average, each question contains 9.03 words, while each answer consists of 10.3 words. Among the training samples, 78.74% have more than 10 relevant reviews. This dataset covers 469,953 products and 38 product categories, and is designed to support product-aware answer generation tasks.
