Imagine have a number of variables what represents explanatory variables of production of sugarcane or soy whatever. For best results in prediction I train my random forest regression model with variables of three year of harvest, so I have:
Variable1 - 2015
Variable2 - 2015
Variable3 - 2015
TargetVariable - 2015
Variable1 - 2016
Variable2 - 2016
Variable3 - 2016
TargetVariable - 2016
Variable1 - 2017
Variable2 - 2017
Variable3 - 2017
TargetVariable - 2017
...
After trainning, how I predict the "TargetVariable - 2018" if I can't "Match Explanatory Variables" of multiples years of harverst? Can I only match 2017 with 2018? Losing the data of other years?
