Hugging Face's logo

Datasets:
sagteam
/
author_profiling

Tasks:
Text Classification
Sub-tasks:
multi-class-classification
multi-label-classification
Languages:
Russian
Size:
10K<n<100K
License:
Dataset card Files Community
1
author_profiling / README.md
naumov-al's picture
naumov-al
First version of the author_profiling dataset.
f961273
preview code
|
raw
history blame
11.4 kB
metadata
annotations_creators:
  - crowdsourced
language_creators:
  - crowdsourced
languages:
  - ru-Ru
licenses:
  - apache-2.0
multilinguality:
  - monolingual
pretty_name: The Corpus for the analysis of author profiling in Russian-language texts.
size_categories:
  - 10K<n<100K
source_datasets:
  - original
task_categories:
  - text-classification
task_ids:
  - multi-class-classification
  - multi-label-classification

Dataset Card for [author_profiling]

Table of Contents

Dataset Description

Dataset Summary

The corpus for the author profiling analysis contains texts in Russian-language which labeled for 5 tasks:

  1. gender -- 13530 texts with the labels, who wrote this: text female or male;
  2. age -- 13530 texts with the labels, how old the person who wrote the text. This is a number from 12 to 80. In addition, for the classification task we added 5 age groups: 1-19; 20-29; 30-39; 40-49; 50+;
  3. age imitation -- 7574 texts, where crowdsource authors is asked to write three texts: a) in their natural manner, b) imitating the style of someone younger, c) imitating the style of someone older;
  4. gender imitation -- 5956 texts, where the crowdsource authors is asked to write texts: in their origin gender and pretending to be the opposite gender;
  5. style imitation -- 5956 texts, where crowdsource authors is asked to write a text on behalf of another person of your own gender, with a distortion of the authors usual style.

Dataset is collected sing the Yandex.Toloka service link.

You can read the data using the following python code:

def load_jsonl(input_path: str) -> list:
    """
    Read list of objects from a JSON lines file.
    """
    data = []
    with open(input_path, 'r', encoding='utf-8') as f:
        for line in f:
            data.append(json.loads(line.rstrip('\n|\r')))
    print('Loaded {} records from {}/n'.format(len(data), input_path))
    
    return data

path_to_file = "./data/train.jsonl"
data = load_jsonl(path_to_file)

Here are some statistics:

  1. For Train file: No. of documents -- 9586 No. of unique texts -- 9586 Text length in characters -- min: 103, max: 12763, mean: 498.1 No. of documents written -- by men: 4767, by women: 4819 No. of unique accounts -- 3054 No. of unique authors -- 3230; men: 1255, women: 1975 Age of the authors -- min: 12, max: 80, mean: 31.1 No. of documents by age group -- 1-19: 734, 20-29: 4477, 30-39: 2604, 40-49: 1063,50+: 708 No. of documents with gender imitation: 1392; without imitation: 2827; not applicable: 5367 No. of documents with age imitation -- younger: 1777; older: 1787; without imitation: 1803; not applicable: 4219 No. of documents with style imitation: 1412; without imitation: 2807; not applicable: 5367.

  2. For Valid file: No. of documents -- 1368 No. of unique texts -- 1368 Text length in characters -- min: 199, max: 2982, mean: 497.9 No. of documents written -- by men: 705, by women: 663 No. of unique accounts -- 437 No. of unique authors -- 461; men: 184, women: 277 Age of the authors -- min: 14, max: 78, mean: 32.4 No. of documents by age group -- 1-19: 88, 20-29: 510, 30-39: 457, 40-49: 242, 50+: 71 No. of documents with gender imitation: 213; without imitation: 425; not applicable: 730 No. of documents with age imitation -- younger: 243; older: 236; without imitation: 251; not applicable: 638 No. of documents with style imitation: 212; without imitation: 426; not applicable: 730.

  3. For Test file: No. of documents -- 2576 No. of unique texts -- 2576 Text length in characters -- min: 200, max: 3262, mean: 503.3 No. of documents written -- by men: 1293, by women: 1283 No. of unique accounts -- 873 No. of unique authors -- 915; men: 357, women: 558 Age of the authors -- min: 13, max: 71, mean: 30.4 No. of documents by age group -- 1-19: 253, 20-29: 1163, 30-39: 713, 40-49: 292, 50+: 155 No. of documents with gender imitation: 356; without imitation: 743; not applicable: 1477 No. of documents with age imitation -- younger: 497; older: 483; without imitation: 497; not applicable: 1099 No. of documents with style imitation: 371; without imitation: 728; not applicable: 1477.

Supported Tasks and Leaderboards

This dataset is intended for multi-class and multi-label text classification.

The baseline models currently achieve the following F1 metrics scores (table):

=== coming soon ===

Languages

The text in the dataset is in Russian.

Dataset Structure

Data Instances

Each instance is a text in Russian with some author profiling annotations.

An example for an instance from the dataset is shown below:

{
  'id': 'crowdsource_4916',
  'text': 'Ты очень симпатичный, Я давно не с кем не встречалась. Ты мне сильно понравился, ты умный интересный и удивительный, приходи ко мне в гости , у меня есть вкусное вино , и приготовлю вкусный ужин, посидим пообщаемся, узнаем друг друга поближе.',
  'account_id': '996ff96ebe8c0c51116f32bff0a55bf0',
  'author_id': 'author_#504'
  'age': 22,
  'age_group': '20-29',
  'gender': 'male',
  'no_imitation': 0,
  'age_imitation': nan,
  'gender_imitation': 1.0,
  'style_imitation': 0.0,
  'meta': {
    'Unnamed: 0': 4915,
    'age': 22,
    'doc_ind': 2408,
    'gender': 1,
    'imitation_type': 'gender_im',
    'source': 'gender_imit_crowdsource',
    'user_id': '996ff96ebe8c0c51116f32bff0a55bf0',
    'doc_id': 
    'id_gender_imit_cs_4916'
  },
}

Data Fields

Data Fields includes:

  • id -- unique identifier of the sample;

  • text -- authors text written by a crowdsourcing user;

  • author_id -- unique identifier of the user;

  • account_id -- unique identifier of the account (several different people (who know each other) could perform a crowdsourcing task under the same account);

  • age -- age annotations;

  • age_group -- age group annotations;

  • no_imitation -- imitation annotations. Label codes:

    • 0 -- there is some imitation in the text;
    • 1 -- the text is written without any imitation

  • age_imitation -- age imitation annotations. Label codes:

    • 'younger' -- someone younger than the author is imitated in the text;
    • 'older' -- someone older than the author is imitated in the text;
    • 0 -- the text is written without age imitation;
    • nan -- not supported (the text was not written for this task)

  • gender_imitation -- gender imitation annotations. Label codes:

    • 0 -- the text is written without gender imitation;
    • 1 -- the text is written with a gender imitation;
    • nan -- not supported (the text was not written for this task)

  • style_imitation -- style imitation annotations. Label codes:

    • 0 -- the text is written without style imitation;
    • 1 -- the text is written with a style imitation;
    • nan -- not supported (the text was not written for this task).

Data Splits

The dataset includes a set of train/valid/test splits with 9586, 1368 and 2576 texts respectively. The unique authors do not overlap between the splits.

Dataset Creation

Curation Rationale

The formed dataset of examples consists of texts in Russian using a crowdsourcing platform. The created dataset can be used to improve the accuracy of supervised classifiers in author profiling tasks.

Source Data

Initial Data Collection and Normalization

Data was collected from crowdsource platform. Each text was written by the author specifically for the task provided.

Who are the source language producers?

Russian-speaking Yandex.Toloka users.

Annotations

Annotation process

We used a crowdsourcing platform to collect texts. Each respondent is asked to fill a questionnaire including their gender, age and native language.

For age imitation task the respondents are to choose a topic out of a few suggested, and write three texts on it:

  1. Text in their natural manner;
  2. Text imitating the style of someone younger;
  3. Text imitating the style of someone older.

For gender and style imitation task each author wrote three texts in certain different styles:

  1. Text in the authors natural style;
  2. Text imitating other gender style;
  3. Text in a different style but without gender imitation.

The topics to choose from are the following.

  • An attempt to persuade some arbitrary listener to meet the respondent at their place;
  • A story about some memorable event/acquisition/rumour or whatever else the imaginary listener is supposed to enjoy;
  • A story about oneself or about someone else, aiming to please the listener and win their favour;
  • A description of oneself and one’s potential partner for a dating site;
  • An attempt to persuade an unfamiliar person to come;
  • A negative tour review.

The task does not pass checking and is considered improper work if it contains:

  • Irrelevant answers to the questionnaire;
  • Incoherent jumble of words;
  • Chunks of text borrowed from somewhere else;
  • Texts not conforming to the above list of topics.

Texts checking is performed firstly by automated search for borrowings (by an anti-plagiarism website), and then by manual review of compliance to the task.

Who are the annotators?

Russian-speaking Yandex.Toloka users.

Personal and Sensitive Information

All personal data was anonymized. Each author has been assigned an impersonal, unique identifier.

Considerations for Using the Data

Social Impact of Dataset

[Needs More Information]

Discussion of Biases

[Needs More Information]

Other Known Limitations

[Needs More Information]

Additional Information

Dataset Curators

Researchers at AI technology lab at NRC "Kurchatov Institute". See the website.

Licensing Information

Apache License 2.0.

Citation Information

If you have found our results helpful in your work, feel free to cite our publication. 

Citation Information coming soon here.

Contributions

Thanks to @naumov-al for adding this dataset.