What would be really useful and I see most articles fail to do is provide a clear approach on how can someone use such trained model to generate returns. If you manage to do that, you can quite literally take your approach to the bank and start cashing in.

Additionally, I don’t think predicting the prices is very useful. You can get good results with simple moving averages that require no fancy algorithms. There is a reason why people aren’t getting rich by recycling their machine learning models from other projects and then using crypto data to train them. You article caught my attention because literally that’s the first thing I did when I learned about LSTMs. Hahhaa.