Italian Legal Ner

🚀 Italian Legal Named Entity Recognition (NER)

An ELECTRA-based model designed to extract key entities from Italian civil judgments issued by the Corte Suprema di Cassazione.

🚀 Quick Start

This model is trained to extract specific named entities from Italian civil judgments. It's based on the ELECTRA architecture and fine - tuned on a dataset of judgments.

📦 Installation

No installation steps are provided in the original document, so this section is skipped.

✨ Features

• The model is fine - tuned on 9000 judgments from 2016 to 2021 (1500 per year), labeled using a combination of rule - based and manual approaches.
• It can extract various named entities from the text, as shown in the table below.

📚 Documentation

Dataset

The model has been fine - tuned on 9000 judgments spanning from 2016 to 2021 (1500 per year). The data was labeled using a combination of rule - based and manual approaches.

Named Entities

It can be used to extract the following named entities from the text.

💻 Usage Examples

Basic Usage

Here is an example of a judgment text:

la seguente SENTENZA sul ricorso 24817 - 2015 proposto da: ANDREA FORMISANO, elettivamente domiciliato in ROMA VIA S. TOMMASO D'AQUINO 7, presso lo studio dell'avvocato CARLO BORELLO, che lo rappresenta e difende giusta delega in calce; - ricorrente - contro SOGET SPA, CAMERA DI COMMERCIO DI PESCARA; - intimati - avverso la sentenza n. 169/2012 della COMM.TRIB.REG.SEZ.DIST. di PESCARA, depositata il 13/03/2012; udita la relazione della causa svolta nella pubblica udienza del 04/04/2018 dal Consigliere Dott. MILENA BALSAMO; udito il P.M. in persona del Sostituto Procuratore Generale Dott. GIOVANNI GIACALONE che ha concluso per l'inammissibilità in subordine rigetto del ricorso.

The model can be used to extract entities such as:

• RIC: 24817 - 2015
• RCR: ANDREA FORMISANO
• AVV: CARLO BORELLO
• CNS: Dott. MILENA BALSAMO
• PMI: Dott. GIOVANNI GIACALONE
• SNT: The entire judgment text

Advanced Usage

For more complex judgment texts with multiple appeals and parties, the model can still accurately extract the relevant named entities. You can use it in a Python script to process a batch of judgment texts and collect the extracted entities for further analysis.

# Here you can use the model's API to process the text
# Assume there is a function `extract_entities` provided by the model
text = "la seguente SENTENZA sul ricorso 24817 - 2015 ..."
entities = extract_entities(text)
print(entities)

🔧 Technical Details

No technical details are provided in the original document, so this section is skipped.

📄 License

No license information is provided in the original document, so this section is skipped.