Forecast single and multiple time series with machine learning models. Implement backtesting to evaluate models before deployment.

Create lag, window and seasonal features, perform imputation and encoding, extract datetime variables, remove outliers, and more.