README.md

metadata

title: Legalis
emoji: ⚖️
colorFrom: red
colorTo: yellow
sdk: gradio
sdk_version: 3.33.1
app_file: app.py
pinned: false
license: mit

legalis demo

This space is showing a demo of two models I build for a Uni Project about the prediction of court case outcomes. It's based on the similarly named legalis dataset, which contains 2800 court cases from the German Court cases. The dataset is labeled with the outcome of the case (plaintiff won or lost), extracted from the decision in text form (by ChatGPT) and the models are trained to predict the outcome based on the text of the case.

🔗 Links:

RandomForest Model on Huggingface

BERT Model on Huggingface

Labeled Dataset on Huggingface

👨‍💻 Author and Credits:

Author: @LennardZuendorf

• This project is part of a course in HTW Berlin's Business Computing Bachelor Program ("Unternehmenssoftware")