[R] calculate RMSE using bootstrap
Thomas Hoffmann
thomas.hoffmann at uni-bonn.de
Mon Oct 15 17:18:24 CEST 2012
Dear all,
I need to calculate confidence intervals of a and b and root mean square
error (RMSE) of a power law given by
Y = a X^b
I calculate the confidence intervals by:
reg.ln = function(tab.xy, indices){
tab.xy = storage[indices,]
res.lm = lm(log(y)~log(x), data=tab.xy)
coefficients(res.lm)
}
xy.boot = boot(tab.xy, reg.ln, 2000)
boot.ci(xy.boot)
where tab.xy is a dataframe containing x and y.
Is there any way to calculate the mean RMSE of the 2000 fitted models,
e.g. identify the data which are not used in the random sub-sample
generated by boot() and calculate the RMSE of the fitted models.
Kind regards for any help
Thomas
More information about the R-help
mailing list