I\'m stuck at back-transforming a time-series which I transformed through both a ln and sqrt transformation in order to get a more accurate forecast.
This is the forecast