---
license: apache-2.0
dataset_info:
features:
- name: url
dtype: string
- name: title
dtype: string
- name: doc
dtype: string
- name: summary
dtype: string
- name: materia
dtype: string
splits:
- name: train
num_bytes: 25541563
num_examples: 792
- name: validation
num_bytes: 2932410
num_examples: 88
- name: test
num_bytes: 6870636
num_examples: 221
download_size: 18051772
dataset_size: 35344609
task_categories:
- summarization
- text-classification
language:
- it
tags:
- legal
pretty_name: ita_casehold
size_categories:
- n<1K
---
# ITA-CASEHOLD
## Dataset Summary
- This dataset contains the data used in the research of the ITA-CASEHOLD model, an extractive summarization model to extract holdings from Italian Legal Administrative documents.
- The research paper titled 'Legal Holding Extraction from Italian Case Documents using Italian-LEGAL-BERT Text Summarization' is accepted for ICAIL 23.
- It consists of 1101 pairs of judgments and their official holdings between the years 2019 and 2022 from the archives of [Italian Administrative Justice](https://www.giustizia-amministrativa.it/it/web/guest/massime).
- The Administrative Justice system in Italy covers a wide range of issues, including public contracts, environmental protection, public services, immigration, taxes, and compensation for damages caused by the State
### Download the dataset
To download the dataset, use the following lines:
from datasets import load_dataset
dataset = load_dataset("itacasehold/itacasehold")
To split the train, test, and validation dataset, use
dataset = load_dataset("itacasehold/itacasehold", split = 'train')
### Supported Tasks and Leaderboards
Summarization, Multi-class Text classification
### Languages
Italian
### Data Fields
The dataset consists of
- **URL**: link to the document
- **Document**: The document
- **Summary**: The holding of the document
- **Materia** : Legal subject
- **Title** : Title of the document
### Data Splits
- **Train** : 792
- **Validatio** : 88
- **Test** : 221
### Source Data
The data is collected from ['Judicial Administration site'](https://www.giustizia-amministrativa.it/it/web/guest/massime).
### Social Impact of Dataset
Legal holdings are considered the most essential part of a legal decision because they summarize it without going into the merits of the specific case, establish a legal principle and set a legal precedent.
The holdings writing is carried out by legal experts who, starting from a judgment, set out the applied principle of law in a clear, precise, and concise manner.
We approached the problem of extracting legal holdings as an Extractive text summarization task.
This Dataset addresses the Legal holding Extraction topic and so far the first and the only one present in the Italian language.
This dataset contributes to Summarization in the Italian language and Summarization tasks in Legal domains.
Apart from this, the Dataset can also be used as a multi-class text classification task utilizing legal subjects.
### Dataset Limitation
This Dataset specifically focuses on the Italian Legal domain, and it is only in Italian. The documents are only from the period of 2019-2022.
## Additional Information
### Dataset Curators
The Dataset was curated by researchers from Scoula Superiore Sant'Anna as a part of the project ['Guistizia Agile (Agile Justice)'](https://www.unitus.it/it/unitus/mappatura-della-ricerca/articolo/giustizia-agile) funded by the Italian Ministry of Justice.
### Licensing Information
The data sets are distributed under the `Apache 2.0` License. More information about the terms of use of the original data sets is listed [here](https://www.apache.org/licenses/LICENSE-2.0).
### Citation Information
If you use this dataset then, please, cite the following paper:
@inproceedings{10.1145/3594536.3595177,
author = {Licari, Daniele and Bushipaka, Praveen and Marino, Gabriele and Comand\'{e}, Giovanni and Cucinotta, Tommaso},
title = {Legal Holding Extraction from Italian Case Documents using Italian-LEGAL-BERT Text Summarization},
year = {2023},
isbn = {9798400701979},
publisher = {Association for Computing Machinery},
address = {New York, NY, USA},
url = {https://doi.org/10.1145/3594536.3595177},
doi = {10.1145/3594536.3595177},
abstract = {Legal holdings are used in Italy as a critical component of the legal system, serving to establish legal precedents, provide guidance for future legal decisions, and ensure consistency and predictability in the interpretation and application of the law. They are written by domain experts who describe in a clear and concise manner the principle of law applied in the judgments.We introduce a legal holding extraction method based on Italian-LEGAL-BERT to automatically extract legal holdings from Italian cases. In addition, we present ITA-CaseHold, a benchmark dataset for Italian legal summarization. We conducted several experiments using this dataset, as a valuable baseline for future research on this topic.},
booktitle = {Proceedings of the Nineteenth International Conference on Artificial Intelligence and Law},
pages = {148–156},
numpages = {9},
keywords = {Italian-LEGAL-BERT, Holding Extraction, Extractive Text Summarization, Benchmark Dataset},
location = {<conf-loc>, <city>Braga</city>, <country>Portugal</country>, </conf-loc>},
series = {ICAIL '23}
}
