How Do You Do A Nonlinear Least-Squares Estimation In Stata?
Sometimes when we look at data in 2 dimensions the relationship we see reminds us of a certain kind of function. Maybe it’s a polynomial or exponential function for example. We use a technique called nonlinear least-squares estimation to get formulas that fit the data. In this eHow I will do just this with Stata. Say we have a data set that reminds us of a certain kind of function like this one. That data set clearly looks like it is being generated by an exponential function, but how do we find that function? We need to use a non-linear least squares estimation. Go to the Statistics menu then go to Linear Models and Related then go to Other and click Non-linear least squares. You will get this screen. Since we think this data set is exponential, we click “use a preprogrammed substitutable” expression and select an exponential function. Insert the dependent and independent variable and click OK. We get this output. The R^2 is .75, so this is a great fit! The coefficients at the bottom
