I'm using the package 'locfit' for nonparametric regression. This package contains the function 'scb' to compute simultaneous confidence bands. The variance of the data is unknown. Up to now I compute a fit with 'locfit'. Afterwards an estimate of the residual variance is computed by the function 'rv'. The weights in the 'scb'-function are set 1/sigma^2 to compute the confidence band. Is this procedure correct or is there any other way to compute confidence bands with unknown variance? Thanks very much for any help you can offer. Michael G??lger
Apologies for coming to this so late... Variance is rarely known in real life data. You should really consult the book `Local Regression and Likelihood' by Prof. Loader for the details on simultaneous confidence bands. `Locfit' is the support software for that book. Andy> From: Michael G??lger > > I'm using the package 'locfit' for nonparametric regression. > This package > contains the function 'scb' to compute simultaneous confidence bands. > The variance of the data is unknown. Up to now I compute a fit with > 'locfit'. Afterwards an estimate of the residual variance is > computed by the > function 'rv'. The weights in the 'scb'-function are set 1/sigma^2 to > compute the confidence band. > Is this procedure correct or is there any other way to > compute confidence > bands with unknown variance? > > Thanks very much for any help you can offer. > > Michael G??lger > > ______________________________________________ > R-help at stat.math.ethz.ch mailing list > https://stat.ethz.ch/mailman/listinfo/r-help > PLEASE do read the posting guide! > http://www.R-project.org/posting-guide.html > >