Hi All, This is more like some general view. How good is using Deep reinforcement learning for trading. How much successful it could be provided I do add more features other than the ones mentioned in the course.
I have been doing machine learning and deep learning courses in quantra and working on deep neural network models seperately. I got around 65 + accuracy for one of my model. However I would appreciate some success stories or some advice on how effective I can be as a retail trader setting up individual desk for trading using DQN or LSTM models.
Thanks in advance!
Hi Vignesh,
Ideally, the deep reinforcement learning model works best if the input features selected are good in predicting the price movement. Simply, adding more features do not result in more accuracy.
Regarding your accuracy, I believe it is a fairly decent strategy, as even the best strategies do not work on very high accuracy levels. If you take your existing 65+ accuracy strategy and couple it with a good risk management trading rule which implements a good risk-reward ratio, then you should be able to devise a fairly decent trading system.
Hope this helps!