[R] help in M-estimator by R

[S Ellison]

> -----Original Message-----
> i have to do something in robust regression by R programm , 
> and i have some problems as following:
> 
> *the first :*
Afte loading MAS, type 'hubers' (without brackets) and look at the code. That should give you a clue.

> new psi ( psi.a=r/(1+(r/k)^2)) ,k=1.345 and get on new 
> coefficients to the
> model, and, Is all the arguments are necessary ? " i want short code"
> 
> again ... *how can i do this ... *

Look at psi.huber (see above for how) and duplicate the code, substituting your calculation. Then look at the 'psi' argument in ?rlm



*******************************************************************
This email and any attachments are confidential. Any use...{{dropped:8}}