How to do nonparametric regression in Eviews 9?

[timbod2003]

If I want to do nonparametric regression in Eviews, say Local Constant Least Squares or Local Linear Least Squares, with different types of kernels and bandwidth selection, how would I go about doing this? I have searched through the Eviews guide and cannot find anything regarding this.

[EViews Gareth]

EViews doesn't really offer nonparametric regression, other than for drawing fit lines on scatter plots.

[timbod2003]

Ok thanks. Is this something that is in consideration for future releases?

[timbod2003]

Gareth,

When we fit the line to a scatter plot, is there anyway via coding or commands to store the value of the line or equation 13.6 (from Eveiws 9 guide) and/or calculate derivatives at specific points?

[EViews Gareth]

Nope.

[timbod2003]

ok thanks

[EViews Glenn]

Not quite true. You can save the fitted points from the nonparametric fit using Proc/Make Distribution Plot Data... from the group object menu, and use these data to calculate things like derivatives as desired. For the command form, look at the group proc distdata.

[EViews Gareth]

Good point