R: glm(...,family=poisson) plot confidence and prediction intervals I could not find many information on how to plot a confidence interval and a prediction interval for a Poisson regression (for example with glm()).
What are some ways to calculate such intervals (in general and especially for poisson regression)? Would Bootstrap be a good idea?
Thank you for the help.
And sorry about the edit. It was maybe more of a R question before.
 A: R's predict ought to be able to do a confidence interval for a GLM but definitely won't do a prediction interval -- there's an underlying statistical issue here, which I'll discuss in this answer.
i) For some GLMs it doesn't make sense to even try to do a PI - consider a logistic regression with 0/1 responses, and imagine you want say a 95% PI. Anywhere that E(Y) is not very close to 0 or 1, a prediction interval will have to include all of 0 to 1, and when E(Y) is very close to 0 or 1, the interval degenerates to just a point.
ii) for many other GLMs there's no ready analytic prediction interval. For example, there's generally no pivotal quantity for the prediction like there is in the normal case. The Poisson is among those. 
A number of papers have looked at ways to get at approximate prediction intervals for some cases, and there's also options like bootstrapping, but because of issues like those I've mentioned, there's no prediction interval in R's GLM.
[To produce a CI for the mean, see ?predict.glm; predict(fit, type="response", se.fit=TRUE) will give the mean and and standard 
error, which can be used to get an approximate (asympotic) interval. Alternatively, you could use the default type="link" and transform an interval for that. ]
