NZDJPY trend prediction using Deep Learning with TensorFlow

NZDJPY was my favorite pair when I learned forex and became a young private trader in 2013. At least better than simply guessing, Deep Learning is a powerful tool that solves many problems in the financial sphere, it provides better and objective vision about the trend and the evolution of a given market based on data rather than a psychological feeling.

Psychology in trading is all about managing the trader’s emotions because it is one of the primary causes of losses. That’s why using Deep Learning is useful.

The model

  • In blue : the point dataset used to train the model
  • In Red : the actual values
  • In Orange : the predicted values

This is a basic model that could be optimized and enhanced using a multi-input model that takes into consideration the values of technical indicators such as Moving average (MA), Exponential moving average (EMA), Stochastic oscillator, Moving average convergence divergence (MACD). But also macroeconomic indicators!

Code repository (and the jupyter notebook) :

Open chat
Powered by