Page 1 of 1

Gompertz with 5 variables

Posted: Thu Nov 15, 2012 5:47 am
by Qmars
Hello,
I wonder whether is it possible to write a small piece of program to set a Gompertz curve with 5 variables?
The 5 variables are:
1- Starting point value
2- Time to reach MAx value
3- Max Value itself
4- Time to reach Min value
5- Min Value itself.

Cheers :D

Qmars

Re: Gompertz with 5 variables

Posted: Mon Dec 31, 2012 12:48 pm
by DannyBachman
I've also been working on estimating a Gompertz function, although with only three parameters (I think that's the word you want, not variables). You should be able to just specify the function in the "estimate" box of the equation object. For example, where x is your series:

x =c(1)*(exp(c(2)*exp(c(3)*(@trend))))

I've had some trouble getting the non-linear estimation routine to converge unless I make an assumption about the maximum value, c(1). The min value here is assumed to be zero, and that's where it starts, hence only three parameters. If your functional form allows for more parameters---I haven't seen this, but am no expert---you should be able to type it in that way. \

I would be interested in knowing what you are using the Gompertz function for. You are welcome to PM me on this system.

Re: Gompertz with 5 variables

Posted: Wed Jan 02, 2013 12:40 am
by trubador
You may also find it useful to try GBASS add-in, which estimates another type of diffusion model known as Generalized BASS.

Re: Gompertz with 5 variables

Posted: Wed Apr 24, 2013 6:59 am
by Qmars
Hello,
Thanks for your reply. I work in the telecom sector and the reason I need a Gompterz with 5 parameters is due to the fact that telecom technologies have a life cycle therefore in order to depict the trajectory of those technologies decline we need to estimate a) duration of decline (deflection) in terms of time span and b) value of the end point (rundown) in terms of mkt share %. or absolute number. If you would like to discuss more about this topic please contact me on qs44@hotmail.co.uk.

Best,

Q