Tom Wenseleers
Ha scam inherits from gam, so I think the methods for both would be the same. I would indeed use the inbuilt se.fit argument to estimate standard errors for these...
Oh yes and I also remember that Ben Bolker outlined a method that he referred to as "population prediction intervals" in a Chapter of his book, http://ms.mcmaster.ca/~bolker/emdbook/chap7A.pdf, which applies to...
Yes nlme::gls is the common one; there is also MASS:lm.gls, which has a predict method with a SE.fit argument. I noticed this paper where they calculated prediction intervals for a...
Definitely interested in this too! Lifting some code from the propagate package could be the easiest...
Many thanks - yes that will do it, perfect! Could still be nice though to formally define AIC and BIC methods... cheers & thanks again! Ha and would you know...
For the gaussian case (together with glmnet) I was using the formulae used here: https://github.com/gabrielrvsc/HDeconometrics/blob/master/R/ic.glmnet.R As degrees of freedom it is using the effective degrees of freedom of the LASSO,...
Ha yes sorry that's what I meant! The calculation of the log likelihood in `ncvreg `is given in function ``` logLik.ncvreg
For me `BIC `in `ncvreg `for variable selection works fine though: ``` fit2
The correct link & inverse link functions for multinomial (both for nnet::multinom and mclogit::mblogit models - the latter doesn't seem currently supported) (assuming mu and eta would be predictions on...