- On 4 Dec 1998 at 11:13:58, Azucena Aldaz (aaldaz.-a-.unav.es) sent the message

Back to the Top

I am working with Kinetica software from Innaphase. I have created my

own model usin the Designer optino. Now, I want to use my own weighting

scheme. I heve read the manual and it says that it is possible, but I do

not how to do it.

Is anyone working with that program? Could anyone help me to do that?

Thank you very much. - On 10 Dec 1998 at 11:23:19, Peter Ryan (113204.1277.aaa.compuserve.com) sent the message

Back to the Top

Dear Dr. Aldaz,

Please find below a soft coded method that will resolve your question

regarding the user defined weighting scheme when performing PK Modeling in

Kinetica. A full template containing the structure and embedded model has

been sent to your direct email address so that you may run the example with

your own data.

Please do not hesitate to contact me if you have any further questions.

Best regards,

Bertrand Roudier, Ph.D.

InnaPhase Corporation

' Two compartment model with user defined error variance.

' The W column contains the computed variance of the residuals.

' The minimised criteria will be the sum of (Yc(i)-Yobs(i))*(Yc(i)-Yobs(i))

/ W(i)

' Bertrand Roudier, December 12 1999

Dim A As Parameter

Dim alpha As Parameter

Dim B As Parameter

Dim beta As Parameter

Dim X As InputColumn

Dim Yc As ComputedColumn

Dim Yobs As ColumnToFit

dim W As ComputedWeightColumn

Sub FitTwoComp()

For i = 1 To X_count

If x_status(i) = 0 And Yobs_status(i) = 0 Then

Yc(i) = A * Exp(-alpha * X(i)) + B * Exp(-beta * X(i))

W(i) = Yc(i) * 10

End If

Next i

Yc_unit = Yobs_unit

W_unit = "(" + Yc_unit + ")^2"

End Sub - On 11 Dec 1998 at 10:30:29, "Nick Holford" (n.holford.aaa.auckland.ac.nz) sent the message

Back to the Top

> Yc(i) = A * Exp(-alpha * X(i)) + B * Exp(-beta * X(i))

> W(i) = Yc(i) * 10

Huh? Why multiply by 10?

This looks like Iteratively Reweighted Least Squares.

Is W(i) supposed to be proportional to the inverse of the variance of the

residual error in Yc? If so then this model is for Poisson error. A more

common error model is the constant CV (aka proportional) model which would

use W(i)=Yc(i)*Yc(i) if the W(i) are really the inverse of the variances.

What happens if Yc is 0?

Can you estimate parameters of the W model e.g. using Extended Least

Squares?

--

Nick Holford, Dept Neurology,L226

OHSU,3181 SW Sam Jackson Park Rd,Portland,OR 97201,USA

email:n.holford.aaa.auckland.ac.nz tel:+1(503)494-7228 fax:494-7242

http://www.phm.auckland.ac.nz/Staff/NHolford/nholford.htm

Want to post a follow-up message on this topic? If this link does not work with your browser send a follow-up message to PharmPK@boomer.org with "PK modelling - Kinetica" as the subject

PharmPK Discussion List Archive Index page

Copyright 1995-2010 David W. A. Bourne (david@boomer.org)