The Basic idea for this Trader is to look into a square of 32x32
Obs Space of past Renko Graph and trade a given stock action: Buy | Sell | Hold
.
This is based off implementation of NEAT from neat-python
and also uses the pytorch version of NEAT
from UBER
for the neural network RNN
.
This project is gonna be superseded by the use of Weight Agnostic Neural Networks
Version of NEAT
from Google DeepMind and also the DeepGA
from Uber
.