1
//! mean absolute percent error. do we need this? I found it in a medium article https://towardsdatascience.com/the-complete-guide-to-time-series-forecasting-using-sklearn-pandas-and-numpy-7694c90e45c1