ellisp / forecastHybrid

Convenient functions for ensemble forecasts in R combining approaches from the {forecast} package
GNU General Public License v3.0
79 stars 23 forks source link

cvts vignette #66

Closed dashaub closed 6 years ago

dashaub commented 7 years ago

Add a cvts() vignette. This might have to wait while we work out the new API to handle xreg.

dashaub commented 6 years ago

added 70d72629755da8ab8b6843fd66b9deea8d3c3f44

russellcameronthomas commented 6 years ago

In the last example of this vignette, you wrote (emphasis added):

As a final example, suppose we foolish want to implement our own version of naive() for performing cross validation.

I think I understand what you are trying to convey, but it is poorly worded for most readers, especially anyone who doesn't read English as a first language. Nobody will "foolishly want to implement their own version of naive()".

May I suggest this instead:

As a final example, I want to illustrate how you can write a custom model function (for the FUN parameter) and a custom forecast function (for the FCFUNparameter) to cross validate nearly any model. For illustration purposes only, I will implement a naive() model, where the forecast is simply a repetition of the last actual value in the time series.