Attention-based CNN-LSTM and XGBoost hybrid model for stock prediction

Requirements

The code has been tested running under Python 3.7.4, with the following packages and their dependencies installed:

numpy
scikit-learn
statsmodels
pandas
tensorflow
keras
xgboost

matplotlib (for plotting)
nvidia-tensorrt (for cuda (GPU) tensorflow)

The stock data used in this repository was downloaded from TuShare. The stock data on TuShare are with public availability.

Usage

Firstly, run ARIMA.py for pre-processing step by ARIMA model. Then, run the neural network or XGBoost models.

• Run LSTM.py for the single-layer LSTM, multi-layer LSTM, and bidirectional LSTM models.
• Run XGBoost.py for the XGBoost model.
• Run Main.py for our proposed Attention-based CNN-LSTM and XGBoost hybrid model.

Citation

@article{shi2022attclx,
    author={Zhuangwei Shi and Yang Hu and Guangliang Mo and Jian Wu},
    title={Attention-based CNN-LSTM and XGBoost hybrid model for stock prediction},
    journal={arXiv preprint arXiv:2204.02623},
    year={2022},
}