Nowadays, there are tones of analysts that claim that they can predict stock prices with high accuracy days in advance.

However,

• Which method for time series forecasting performs better or achieves a lower mean squared error in critical events?
• Which method performs better with a couple of days in advance prediction? Recurrent Neural Networks vs Moving Averages
• Which is the optimal window size to optimise the forecast?
• Could we develop a method to predict bubbles in the stock market as well as a crisis? blog post
• Can we predict a crisis with several days in advance by understanding the patterns on the stock market just before it happens?

In the content of this repository, these topics will be addressed with the historical data of the S&P 500 Index since 1950.