Hello,
Not sure people have done this on Quantopian. When we train machine learning models, there are a few hyper parameters we need to learn during in sample period. And then using the learned best hyper parameters to learn model weights during out of sample prediction.
So the difficult parts is to hook this whole in sample hyper parameter learning and backtesting plus the out of sample prediction together. This sounds a pretty challenging task. Would appreciate if anyone could direct me to sources or examples done by someone on either Quantopian or other backtesting tools.
Thanks,
CQ