Section 15, Unit 4

data['return_threshold'] = 0.125 * <br />
    np.sqrt(feature_window)*data.daily_returns.rolling(feature_window).std()



In this formula, where's the 0.125 coming from? 

Hello Alex,



So, the value 0.125 is just to create a scaled variant or function of the rolling standard deviation. We took that value based on the data we used for illustration in the notebook. It's perfectly fine to take other values as well based on summary statistics or educated guesses.