paulpall/legalese-sentences_estonian

Estonian Legalese Corpus

Dataset Summary

The Estonian Legal Texts dataset is a collection of legal documents extracted from the Estonian National Corpus. It is tailored for Natural Language Processing (NLP) tasks, particularly those involving the Estonian language. The dataset contains legal texts such as acts, regulations, and legal proceedings, making it an essential resource for developing language models, text classification systems, and other NLP tools for Estonian.

Languages

• Primary Language: Estonian (et)

Dataset Structure

• Format: Plain text
• Documents: The dataset includes various types of legal texts, such as acts, regulations, and court proceedings.
• Annotations: The texts are not annotated, providing raw legal text for analysis and model training.

Data Fields

• text: The raw legal text from the Estonian legal documents.

Data Splits

This dataset has not been pre-split into training, validation, or test sets. Users can create splits based on their specific use case.

Dataset Creation

Source Data

• Origin: The legal texts are sourced from the Estonian National Corpus, which includes a wide range of text types, with legal documents forming a specific subset.
• Collection Process: Legal texts were filtered and extracted from the Estonian National Corpus to create this dataset. The focus was on maintaining the context and integrity of the legal language.

Preprocessing

• Cleaning: The text data was cleaned to remove extraneous characters and ensure consistent formatting. Non-legal content was removed, and the documents were kept in their original Estonian language without additional annotation or translation.

Uses

Primary Use Case

• Grammatical Error Correction (GEC) in Estonian: Useful for training and evaluating GEC models for Estonian.
• Legal Text Processing: Ideal for tasks such as legal document classification, named entity recognition (NER), and language modeling in the legal domain.

Potential Applications

• Legal Document Classification: Classifying legal documents by type, topic, or domain.
• Named Entity Recognition (NER): Identifying legal entities within the texts.
• Language Modeling: Developing models that can understand or generate legal Estonian text.
• Translation Models: Creating models to translate Estonian legal texts to other languages.

Ethical Considerations

• Data Privacy: The texts are from publicly available legal documents. Nevertheless, the accuracy and context in which these texts are used should be carefully considered.
• Bias and Fairness: The dataset may reflect biases specific to the Estonian legal system and culture, and care should be taken when applying models trained on this data to other legal contexts or languages.

Citation

If you use this dataset in your research or project, please cite it as follows:

@dataset{estonian_legal_texts_2024,
title={Estonian Legal Texts from the Estonian National Corpus},
author={Institute of Estonian Language},
year={2024},
url={https://example.com/estonian-legal-texts}
}

Licence

The original source just states that "The corpus is free for non-commercial use."

Dataset Curators

The dataset was curated by researchers at the Institute of Estonian Language, in collaboration with the Estonian National Corpus project.

Contact Information

For any inquiries or issues related to this dataset, please contact the dataset curators at [info@example.com].

---

This dataset card is formatted according to the Hugging Face Dataset Card Template.