- This repository houses the datasets/resources used in our paper ChatGPT Informed Graph Neural Network for Stock Movement Prediction.
- This research introduces a novel framework that leverages ChatGPT's graph inference capabilities to enhance Graph Neural Networks (GNN).
- Our model shows superior performance in both (1) stock movement prediction and (2) portfolio construction.
- Dive in to find the datasets, code samples, and more!
- 📄 Paper: [Link to the paper]
The data can be found in the Data folder, which contains two files:
ticker_train_data.json: This file holds the data utilized for training and validation of our model.ticker_test_data.json: This file contains the data used for model evaluation.
To load the data, you can start with 4 lines of code:
import pandas as pd
import json
train_data = pd.read_json('./Data/ticker_train_data.json')
test_data = pd.read_json('./Data/ticker_test_data.json')
The Affected Companies column provides two key insights:
- Companies that ChatGPT predicts will be influenced by the financial news.
- The sentiment indicating the nature of the impact on these companies (e.g., positive or negative).
For a deeper exploration of the data, please feel free to check the data_checking.ipynb.
We encourage collaboration and use of this dataset for further advancements in stock prediction using deep learning. If you find this resource useful, kindly cite our paper. Happy researching!
@article{chen2023chatgpt,
title={ChatGPT Informed Graph Neural Network for Stock Movement Prediction},
author={Chen, Zihan and Zheng, Lei Nico and Lu, Cheng and Yuan, Jialu and Zhu, Di},
journal={arXiv preprint arXiv:2306.03763},
year={2023}
}