Changes in version 1.0.5 (2024-02-15)                  

  - Now, if the lags are automatically chosen, it is not allowed to use
    only one autoregressive lag with the additive ormultiplicative
    transformation (it does not make sense)

                 Changes in version 1.0.4 (2023-03-24)                  

  - autoplot.grnnForecast has been modified to comply with CRAN

                        Changes in version 1.0.3                        

  - citation information has been added

                 Changes in version 1.0.2 (2022-05-30)                  

  - References to the paper describing this package have been added in
    the documentation

                 Changes in version 1.0.1 (2021-04-22)                  

  - The default transformation is now additive instead of
    multiplicative.

                 Changes in version 1.0.0 (2021-03-08)                  

  - Automatic computation of smoothing parameter can be done with
    rolling origin or fixed origin (less accurate, but faster).

  - Lags can be selected using forward selection or backward elimination
    (feature selection techniques)

  - Time series cannot be scaled

  - The model is built using faster Rcpp code

  - An optional transformation to the training samples has been added.
    It improves forecast accuracy for time series with a trend

                        Changes in version 0.2.0                        

  - Uses Rcpp for speeding up computation of forecasts.

                 Changes in version 0.1.0 (2019-11-25)                  

  - This is the first release of this package.