Dmitry Lukovkin, CEO of
StocksNeural.net, communicates his outlook on the application of Deep Learning to the problem of the stock prices analysis and prediction in the AI section of the “Startup Innovation and investment in Emerging Europe" report by EWDN.
Some highlights:
Straight-forward approach using RNN for future prices predictions seems to be inefficient
Customized CNN-based architectures show better results
Deep Reinforcement Learning applied directly to the trading decision making could be a game-changer
Industry adoption of Deep Learning-powered trading will rise both on traditional and crypto markets.
You can download the report here -
http://cee.ewdn.com/.