Stock Prediction [7]: Creating the Model

Ethan Shaotran

Former C# Programmer for multiple renowned iOS applications, and now Main Author for Energize.AI

You may also like...

2 Responses

  1. Hiro says:

    How did you decide on the number of layers and neurons? Is it basically trial and error to see what works best?

    • Hi Hiro – Yes, the number of layers is largely based on “trial and error” – there is no golden rule on how many hidden layers a model should have. Generally, more complicated inputs require more layers – keep in mind, though, that larger models are more computationally expensive. Good luck!

What are Your Thoughts?