Coefficient Restrictions for NLS

[rdwest]

Hi,

I am estimating a diffusion model for which the two parameters to be estimated (c(1) and c(2)) must be bounded between 0 and 1. I'm having trouble figuring out how to impose these restrictions on the model.

I am familiar with the command "param" to set starting values for iterative estimation....but I can't seem to impose restrictions (or "bounds", as programs like SAS name them).

Here is the model (where time is the independent variable)

param c(1) 0.005 c(2) 0.005
equation eq1.ls dep = (1+ exp(-(c(1)+c(2))*time))/(1+(c(1)/c(2))*exp(-(c(1)+c(2))*time))

Thank you,
R.

[EViews Gareth]

Search is your friend:

http://forums.eviews.com/viewtopic.php? ... ction#p103