zshicode/attention-clx-stock-prediction

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

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==1.16.5
sklearn==0.21.3
statsmodels==0.10.1
pandas==0.25.1
tensorflow==2.1.0
keras==2.3.1
xgboost==1.5.0

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},
}

attention-clx-stock-prediction's Issues

Discussion: How to use this ML model on other market

This is not to report an issue with this project; but rather a discussion on how to use this model on other financial instruments.
This was a great research! Well done.
It looks possible to port this model on to other financial instruments; e.g. Forex, Indexes, Bonds, Futures etc.
Would hope to know how would you suggest to tune/improve/change/test this ML model please?
Many thanks.

Can't find the training data used in the research paper

Hey Shi, sorry for opening an issue but i couldn't text you directly, i'm interested in your research and i couldn't find the data used for this model since the website that you linked to is in chinese and i couldn't understand
could you please send me the direct link for the data or an excel version

thank you in advance

提问

为什么在代码中并未发现这个（ARIMA_residuals1.csv')部分

What is K in the source code?

what is K in here?
a = Lambda(lambda x: K.mean(x, axis=1), name='dim_reduction')(a)

https://github.com/zshicode/Attention-CLX-stock-prediction/blob/37c471ee207f9f9c2248672f0e4ca03136ddad0c/model.py#LL12C1-L23C1

ARIMA_residuals1.csv

File "D:\installsoft\anaconda3\Lib\site-packages\pandas\io\common.py", line 863, in get_handle
handle = open(
^^^^^
FileNotFoundError: [Errno 2] No such file or directory: './ARIMA_residuals1.csv'

没有这个文件啊