I'm trying to make a plot of the uncertainty interval of in-sample prediction

Y_obs <- read.table("series.txt")
arma <- arima(Y_obs, order = c(2,0,1))
Y_in_sample <- Y$V1 - arma$residuals

If I want to show some sort of interval of the uncertainty, so I could write

std <- sd(Y_in_sample)
int <- c(Y_in_sample - 1.96*std, Y_in_sample - 1.96*std)

but does this would be the confidence interval for the mean, and not really a prediction interval, so I'm not sure if this makes sense. Is there a natural way of making such an uncertainty interval for in-sample prediction? And would this interval be a confidence interval or a prediction interval?

(It seems some literature does not distinguish between the two, so in case there is no distinction between the two in this case, I'd want to know.)


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.