Ridge Regression - Choosing "k"

[Gabriella Miniero]

Hi Forum!
I have to do ridge regression but I'm in trouble with the ridge parameter. I tried to compare RMSE with different values of "k" (0 0.01 0.05...10) and then I put them on a graph in order to choose the k parameter which minimizes the RMSE. But I'm not sure that this procedure is correct, just because an increase of k correspond an increase of the RMSE. This would mean I would choose k = 0 and I think that this criterion makes no sense. What am I doing wrong?

[Scott Merryman]

It would be helpful if you specified the command you are using and even better would be an example.

[Gabriella Miniero]

Yes. I searched the RMSE manually; then, I ran "ridgereg" on Stata by using the "k parameter" chosed like I said before.
I attached a graph that shows the results of my procedure. The question is: is that correct? or Is there any way I can get that on Stata?