sogou_news / README.md
albertvillanova's picture
Convert dataset sizes from base 2 to base 10 in the dataset card (#2)
52964e5
|
raw
history blame
No virus
6.4 kB
metadata
pretty_name: Sogou News
dataset_info:
  features:
    - name: title
      dtype: string
    - name: content
      dtype: string
    - name: label
      dtype:
        class_label:
          names:
            '0': sports
            '1': finance
            '2': entertainment
            '3': automobile
            '4': technology
  splits:
    - name: test
      num_bytes: 168645860
      num_examples: 60000
    - name: train
      num_bytes: 1257931136
      num_examples: 450000
  download_size: 384269937
  dataset_size: 1426576996

Dataset Card for "sogou_news"

Table of Contents

Dataset Description

Dataset Summary

The Sogou News dataset is a mixture of 2,909,551 news articles from the SogouCA and SogouCS news corpora, in 5 categories. The number of training samples selected for each class is 90,000 and testing 12,000. Note that the Chinese characters have been converted to Pinyin. classification labels of the news are determined by their domain names in the URL. For example, the news with URL http://sports.sohu.com is categorized as a sport class.

Supported Tasks and Leaderboards

More Information Needed

Languages

More Information Needed

Dataset Structure

Data Instances

default

  • Size of downloaded dataset files: 384.27 MB
  • Size of the generated dataset: 1.43 GB
  • Total amount of disk used: 1.81 GB

An example of 'train' looks as follows.

{
    "content": "du2 jia1 ti2 go1ng me3i ri4 ba4o jia4 \\n  re4 xia4n :010-64438227\\n  che1 xi2ng ba4o jia4 - cha2 xu2n jie2 guo3 \\n  pi3n pa2i   xi2ng ha4o   jia4 ge2   ji1ng xia1o sha1ng   ri4 qi1   zha1 ka4n ca1n shu4   pi2ng lu4n ",
    "label": 3,
    "title": " da3o ha2ng "
}

Data Fields

The data fields are the same among all splits.

default

  • title: a string feature.
  • content: a string feature.
  • label: a classification label, with possible values including sports (0), finance (1), entertainment (2), automobile (3), technology (4).

Data Splits

name train test
default 450000 60000

Dataset Creation

Curation Rationale

More Information Needed

Source Data

Initial Data Collection and Normalization

More Information Needed

Who are the source language producers?

More Information Needed

Annotations

Annotation process

More Information Needed

Who are the annotators?

More Information Needed

Personal and Sensitive Information

More Information Needed

Considerations for Using the Data

Social Impact of Dataset

More Information Needed

Discussion of Biases

More Information Needed

Other Known Limitations

More Information Needed

Additional Information

Dataset Curators

More Information Needed

Licensing Information

More Information Needed

Citation Information

@misc{zhang2015characterlevel,
    title={Character-level Convolutional Networks for Text Classification},
    author={Xiang Zhang and Junbo Zhao and Yann LeCun},
    year={2015},
    eprint={1509.01626},
    archivePrefix={arXiv},
    primaryClass={cs.LG}
}

Contributions

Thanks to @lhoestq, @mariamabarham, @lewtun, @thomwolf for adding this dataset.