- on Ubuntu 20.04.5
- python 3.8
- numpy 1.24.2
- pandas 1.1.5
- pytorch 1.7.1
- tensorboard 2.14.0
- scipy 1.10.1
- For the above HRL/RL model use .csv file which contains
open,high,low,close,volume,vwap,targetas input. The dataset is of two types i.e.arg.add_arg("env_name", "hs300", "Environment name, can be [hs300, sp500,nifty100,csi800]"). - There are three ways to calculate rewards (a measure of the correlation between mathematical expressions of generated High Frequency Risk Factors and targets) in
class Risk_Factor_Env: IC*, RankIC*, and IR*.
python run_main_hppo_to.py
python run_main_hrl.py
python run_main_ppo.py
