Re: [R] ARIMA models - estimation prediction$se

From: DGR <ana.lourenco.quiterio_at_caixaseguros.pt>
Date: Thu, 30 Sep 2010 12:23:12 +0100

Dear All

Can anyone help me?

fit <- arima(USAccDeaths, order = c(0,1,1),seasonal = list(order=c(0,1,1)))) fit$sigma2
[1] 99346.89

So, the standard error for my first step prediction is sqrt(fit$sigma2)=315.1934 like predict(fit, n.ahead = 6)$se[1]

> predict(fit, n.ahead = 6)

$pred

          Jan Feb Mar Apr May Jun 1979 8336.061 7531.829 8314.644 8616.869 9488.912 9859.757

$se

          Jan Feb Mar Apr May Jun 1979 315.4481 363.0056 405.0168 443.0622 478.0896 510.7203

And now, How can I calculate the standard errors for the 5 steps ahead?

Thanks in advance.

Ana

---

R-help_at_r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code. Received on Thu 30 Sep 2010 - 11:26:04 GMT