PPN

We provide the original implementation for "Cost-Sensitive Portfolio Selection via Deep Reinforcement Learning (IEEE TKDE 2020)". The paper is available [here].
Note that this library is several versions ahead of the article. The main difference is the decision making module, where we develop a new leverage operation that contributes a lot.

Dependencies

Python 2.7
Tensorflow>=1.4.0
Tflearn>=0.3.2
pympler>=0.5
cvxopt>=1.1.9
seaborn>=0.8.1
pandas>=0.20.3

Usage

Dataset

The attached dataset is the S&P500, whose description can be found in the paper.
Please download the dataset [here], and put the dataset file (databse) in the main directory.

Training

There are three steps (Please use this code after downloading the dataset):

go into train_package/1 file, and modify the net_config.json file for parameter setting (one can also construct train_package/2)
vim the main.py file, and set a specific GPU device
python main.py --mode=train --process=1
(logging information: view train_package/1/programlog)

Backtest

python main.py --mode=backtest --algo=1
(--algo could be either the name of traditional method or the index of training folder)

Save and Restore of the Model

The trained weights of the network are saved at train_package/1 named as netfile (including 3 files)

Plotting

python main.py --mode=plot --algos=crp,olmar,1 --labels=crp,olmar,ours
(--algos could be the name of the tdagent algorithms or the index of nnagent)
(--labels is the name of related algorithm that will be shown in the legend)

Present backtest results in a table

python main.py --mode=table --algos=1,olmar,ons --labels=nntrader,olmar,ons

Acknowledgement

This project is constructed based on the open source project:

[PGPortfolio toolbox(https://github.com/ZhengyaoJiang/PGPortfolio)]

It would not have been finished without using the codes from the following open source projects:

Online Portfolio Selection toolbox

Risk Disclaimer

There is always risk of loss in trading. All trading strategies are used at your own risk.

Citation:

If you use this code, please cite:

@article{zhang2020cost,
  title={Cost-sensitive portfolio selection via deep reinforcement learning},
  author={Zhang, Yifan and Zhao, Peilin and Li, Bin and Wu, Qingyao and Huang, Junzhou and Tan, Mingkui},
  journal={IEEE Transactions on Knowledge and Data Engineering},
  year={2020},
  publisher={IEEE}
}

atilimcetin / PPN

PPN