This version is released on the 22–6-2013 and it is the first time that robust (sandwich) standard errors are introduce in gamlss models. Of course those standard errors apply to parametric GAMLSS models only. When non-parametric smoothing terms are used then the (sandwich) standard errors can still be used with caution since they are not yet take into account the variability in those terms.
There are two new functions for sandwich calculations:
- i) get.K() and
- ii) rvcov().
The first is getting the meat (in the sandwich calculations, that is the matrix K) while the second is the robust version of vcov(). Also a new argument “robust” is introduced for the gamlss() methods a) vcov.gamlss() b) summary.gamlss() and c) confint.gamlss().
Some bugs are fixed on the following function
- LR.test() a small amendment for checking whether it is gamlss model or not
- gen.likelihood() a small amendment to make sure that it gets the correct link functions
- getSmo() is introduced for get information on fitted smoothing terms
- lms() a bug on the previous version is corrected and a additional feature is added with the argument”trans.x”. When “trans.x=TRUE” the function looks for a suitable transformation in the x-variable.