A Reservoir Computing (RC) implementation in Mathematica for financial time series prediction. Brute force and genetic algorithm optimization techniques are included for performance tuning.
Sample SP500 as well as foreign exchange data has been scraped and cleaned for use. Minute resolution ticker SPY ticker data is also included for more thorough backtesting.
Some timeframes appear to be very predictable, while others appear to be noise.
A PNL backtester has also been implemented to test different trade signals.
