I am working with a time series dataset using the `fable` package (R). I have fitted several models (e.g., ARIMA) and generated forecasts. The accuracy calculation is resulting in NaN values, and the warning suggests incomplete out-of-sample data.
I am seeking guidance on how to handle this incomplete out-of-sample data issue and successfully calculate accuracy metrics for my time series forecasts. If anyone has encountered a similar problem or has expertise in time series analysis with R, your insights would be greatly appreciated.
accuracy(data.fit) Error in accuracy.default(data.fit) : No accuracy method found for an object of class mdl_dfNo accuracy method found for an object of class tbl_dfNo accuracy method found for an object of class tblNo accuracy method found for an object of class data.frame
how do i solve this error ?
Dear Sachin. You should review every step in such a computation. For all those fcst evaluation criteria to produce NaN's, you should have NaN's generated in the fcst errors, which is really basic stuff.
- Badges
- Science topic