Hi guys, I wanted to share a personal project I working on with a view to launching a product in the near future.
I have built and continue to test, a Machine Learning model that uses years worth of stock/forex/crypto pricing data and various momentum indicators for each. It then attempts to predict the price a Ticker will close at tomorrow.
There are still some weaknesses in the model, but I am seeing some successes in GOLD, BTC & AAPL.
While the model may not predict the closing price to the cent (although it might), what matters is if it predicts the price goes UP or DOWN, by at least a certain %. This then becomes tradable information.
Model is built using Python & SKLearn and backtested using Lumibot and data from YFinance.
If you want to learn more, please let me know. I might post some predictions in the coming days