Search code examples
rforecastingforecast

Why autoplot function don't show the 95% confidence interval, but plot function do?

The autoplot function don't display the 95% confidence interval in graph. It show only 80% ic.

enter image description here

Plot function displays both ic, 80% and 95%.

enter image description here What I'm doing wrong in the autoplot?

I already tried to change the autoplot arguments, but nothing seen to change the output graph.

The data I'm using is a ts (time series) like this:

Time-Series [1:71] from 2012 to 2018: 3.0515 -2.0783 -1.2 -3.2055 -0.0132 ... enter image description here

The model output looks like this: enter image description here

forecasts <- HoltWinters(temp, beta=FALSE, gamma=FALSE, l.start=3.0515) 
forecasts_12 <- forecast(forecasts, h=12) #projeção 12 meses a frente
plot(forecasts_12)
autoplot(forecasts_12)
autoplot(forecasts_12, PI = TRUE,shaded = FALSE, pi.col="#FF3B6F", fcol = "#0000AA", flwd = 0.5, se = FALSE)

I would like to display the confidence interval of 80% and 95% with autoplot function.

Solution

  • I think you have loaded the ggfortify package which has its own autoplot function, and that is masking the autoplot function from the forecast package. I strongly recommend you don't use ggfortify. It doesn't play nicely with other packages.