This repository contains the Yahoo! news dataset in paper Read, Attend and Comment: A Deep Architecture for Automatic News Comment Generation (EMNLP2019). The dataset is available here.
We build this dataset by crawling news articles and the associated comments from Yahoo! News. The side information associated with news including:
- Paragraph. After pre-processing, we retain the paragraph structure of news article.
- Category. There are 31 news categories and the distribution is shown in figure 1.
- Wiki-Entities. The Wikipedia entities mentioned in the news articles are extracted.
- Vote. Each comment has upvote, downvote and abusevote information from news readers.
- Sentiment. Each comment is annotated with POSITIVE, NEGATIVE or NEUTRAL by Yahoo!.
After pre-processing, We randomly sample a training set, a validation set and a test set. Please refer to the paper for more details.
Train | Validation | Test | |
---|---|---|---|
# News | 152,355 | 5,000 | 3,160 |
Avg. # Comments per News | 20.6 | 20.5 | 20.5 |
Avg. #Upvotes per Comment | 31.4 | 30.2 | 32.0 |
Avg. #DownVotes per Comment | 4.8 | 4.8 | 4.9 |
Avg. #AbuseVotes per Comment | 0.05 | 0.05 | 0.05 |
If you use this dataset in your research work, please cite our EMNLP2019 paper.
@inproceedings{yang-etal-2019-read,
title = "Read, Attend and Comment: A Deep Architecture for Automatic News Comment Generation",
author = "Yang, Ze and
Xu, Can and
Wu, Wei and
Li, Zhoujun",
booktitle = "Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing (EMNLP-IJCNLP)",
month = nov,
year = "2019",
address = "Hong Kong, China",
publisher = "Association for Computational Linguistics",
url = "https://www.aclweb.org/anthology/D19-1512",
doi = "10.18653/v1/D19-1512",
pages = "5076--5088",
}