[S] time covariate in nlme

WayWay M. Hlaing (wayway@gate.net)
Tue, 7 Apr 1998 17:12:03 -0400

Dear S+- users:

I am using a gompertz function NLME model where dependent var. is a timed
(BP)+AD0-blood pressure. I like to add a time covariate (e.g. WT+AD0-weight) as an
variable to the model written below:

BP.FUNC +ADw-- function(A, B, C, D, AGE)+AHs- A +- exp(B - exp(C +- D +ACo- (AGE - 6)))+AH0-

model3 +ADw-- nlme(BP +AH4- BP.FUNC(A, B, C, D, AGE),
fixed +AD0- list(A +AH4- SEX+-RACE, B +AH4- SEX+-RACE, C +AH4-., D +AH4- .),
random +AD0- list(A +AH4- ., C +AH4- .),
cluster +AD0- +AH4- ID,
start +AD0- list(fixed +AD0- c( 87.82729, 0, 0, 3.332103, 0, 0,
.1479483, -0.2751992)),
na.action +AD0- na.omit,
control +AD0- list(tolerance +AD0- 0.001, pnls.tolerance +AD0- 0.01, lme.tolerance +AD0-
data +AD0- bpdata, verbose +AD0- T)

My questions:

(1) Is it as simple as adding +ACI-WT+ACI- to argument... fixed+AD0- .. as follows:
fixed+AD0-list(A+AH4-SEX+-RACE+-WT, B+AH4-SEX+-RACE+-WT, C+AH4-., D+AH4-.),
giving an appropriate starting values for WT in start+AD0-list(fixed+AD0-c(...)?

(2) When is it appropriate to use an optional argument: +ACI-serial.covariate+ACI- -
I did not
understand clearly its use after reading from the manual.

Any help +ACY- / or references would be greatly appreciated.

I use win NT splus 4.0 release 3.

Thanks in advance, WayWay M. Hlaing, University of South Florida


This message was distributed by s-news@wubios.wustl.edu. To unsubscribe
send e-mail to s-news-request@wubios.wustl.edu with the BODY of the
message: unsubscribe s-news