Page 1 of 1

Non-linear least squares in Eviews

Posted: Fri Apr 17, 2020 8:51 am
by londonphd
Hello,

I am trying to estimate the following non-linear model via the Estimation Equation menu:

Code: Select all

logy = c(3)+(1/c(2))*log((((eal*x1)^c(1)+(eak*x2)^c(1))^(c(2)/c(1))+(eae*x3)^c(2)))
I selected LS- Least Squares (NLS and ARMA) from the list of methods. Is this the correct option for NLS modes in Eviews?

But I am getting "Insufficeint number of observations" message, though, the data has 215 observations.

Re: Non-linear least squares in Eviews

Posted: Fri Apr 17, 2020 8:53 am
by EViews Gareth
Yes that is the correct menu item.

Make sure the starting values (located in the C vector in the workfile) are sensible, and that you don't have NAs etc... in your data.

Re: Non-linear least squares in Eviews

Posted: Fri Apr 17, 2020 9:23 am
by londonphd
Thanks Gareth, works like a charm.