jrmeier/fast-trade

AutoML Support?

Opened this issue · 0 comments

Is there a way to churn out multiple strategies without breaking a sweat by testing possible variations? It has been said that XGBoost and algorithms like them are useful for trading algorithms.
A secondary thought would be to get TA-Lib to churn out all their indicators based on the time frame provided in multiple granularity (similar to what is done with MA), and deciding on the Boolean output criteria (e.g. whether the performance after X days exceed y%).
Would there be a way to use fast-trade to accelerate the search?