This repository contains the code and data used in our paper titled "Expediting History: Using AI to Uncover Hidden Authorships of SCOTUS Opinions in Real Time". The paper is under review in Science and suggests AI model for Identifying Authorship in SCOTUS Opinions 1994-2023.
You can use the tool to identify author of opinions using Visit SCOTUS_AID
Dataset used to train and evaluate the model: raminass/opinions-94-23
The model reported in the paper and website, is the one trained with 17 justices, code and dependency packages used to collect and build the data train_model.ipynb.
Code used to run multiple models and select the best one select_model.ipynb, also each one of the experimented models provided within m1-m10.ipynb.
The best selected model is the m4.ipynb.
This repository contains the code and data used in our paper titled "Expediting History: Using AI to Uncover Hidden Authorships of SCOTUS Opinions in Real Time". The paper is under review in Science and suggests AI model for Identifying Authorship in SCOTUS Opinions 1994-2023.
You can use the tool to identify author of opinions using Visit SCOTUS_AID
Dataset used to train and evaluate the model: raminass/opinions-94-23
The model reported in the paper and website, is the one trained with 17 justices, code and dependency packages used to collect and build the data train_model.ipynb.
Code used to run multiple models and select the best one select_model.ipynb, also each one of the experimented models provided within m1-m10.ipynb.
The best selected model is the m4.ipynb.
