Closed ztwe closed 1 year ago
Thanks for your interest on TradeMaster. May I ask what is the data structure of the tick data we want to train? What trading tasks are you working on?
The data structure is like bid_5/bid_4/bid_3/bid_2/bid_1、ask_1/ask_2/ask_3/ask_4/ask_5, both for price and order. The training is about chinese future.
You may take a look at this tutorial https://github.com/TradeMaster-NTU/TradeMaster/blob/1.0.0/tutorial/Tutorial6_DDQN.ipynb. I think it is using data with similar structures on Crypto trading.
Besides data structure, could the environment support the trade which less than 1min, like buy and then sell within 1min.
In addition, could the models and environment accept tick data(...bid_2/bid_1、ask_1/ask_2/...) and OHLCV data simultaneously?
As far as I know, the current version is using technical indicators for states. I think it is possible to modify it to accept tick and OHLCV together. But intuitively, this may not be a very good way because it can be hard to train RL agents with good results using only raw tick and OHLCV data based on my previous experience.
Thank you!
Can trademaster train using tick data?
If not, how to modify the project?