Intraday trading is a fundamental quantitative trading task, where traders actively long/short one pre-selected financial asset within the same trading day to maximize future profit.
DeepScalper use deep q network to optimize the reward sum got from reinforcement learning where a hindsight reward is used to capture the long-term porfit trends and embedding from both micro-level and macro-level market information.
In this notebook, we implement the training and testing process of DeepScalper based on the TradeMaster framework.
And this is the script for training and testing.