Semipar regression with more than two nonlinearly variables

[Oscar Vidal]

Hello,

I'm trying to ran a semipar. My problem is that the semipar command only permit that one variable entered to the model nonlinearly (the nonparametric component), but I need more than two variable nonlinearly.
Does anyone know how to do it? I've searched for others commands but I have not found a way to solve it.

[Nick Cox]

semipar is from the Stata Journal, as you're asked to explain.

SJ-12-4 st0278 . Robinson's square root of N consistent semipar. reg. estim.
(help semipar if installed) . . . . . . . . V. Verardi and N. Debarsy
Q4/12 SJ 12(4):726--735
presents Robinson's double residual semiparametric regression
estimator and Hardle and Mammen's specification test

As for the question

search mvrs

for one possibility. Depending on your definition of "semi-parametric" it's arguable that mfp qualifies too.

[Oscar Vidal]

Hello Nick, thanks for your answer. The problem is that I have a model:

yi = θ0 + xi*θ + f(z) + εi

where yi is the value taken by the dependent variable for individual i, xi is the row vector of characteristics of individual i, θ0 is a constant term, z is an explanatory variable that enters the equation nonlinearly according to a nonbinding function f(z) (the non parametric component). I need that the f(z) function may depend of more variables, e.g. f(z, w, h).

[Nick Cox]

I see. Well, Stata is programmable and the answer to many questions is that you may need to program that yourself. But watch out for other answers. I am not at all confident that I know all the commands in this territory.

[Ratih Twi]

Hi Oscar, have you find the answer for your problem?
Because I have the same problem with you and still don't know what to do with that.
Would you please sharing your solution if in case you already have it?
Thank you

[FernandoRios]

Hi Raith
There isnt a premade solution for that, but if you have Stata 16, it is relatively straight forward with the newest npregress series.
If you have Stata 15, you can also use npregress kernel to do what you need to do. which version of stata do you have?
Fernando