Message-ID: <60129.201.9.73.157.1134530569.squirrel@www.uefs.br>
Date: 2005-12-14T03:22:49Z
From: Evandro do Nascimento Silva
Subject: Memory shortage running Repeated Measures (nlme)
In-Reply-To: <mailman.234.1134538211.31625.r-help@stat.math.ethz.ch>
Dear group,
I tried to run a Repeated Mesures Anova for Mixed effects model and I got
a warnning after entering the model specification saying: "Reached total
allocation of 254Mb: see help(memory.size)".
here is part of the log:
***********************************************************
> aphids<-read.table("aphid.txt",header=T)
> attach(aphids)
> names(aphids)
[1] "site" "time" "shade" "treatp" "aphid"
> library(nlme)
> aphids<-groupedData(aphid~time|site/shade/treatp,data=aphids,outer=~shade*treatp)
> model<-lme(aphid~shade*treatp,random=~time|site)
Error in logLik.lmeStructInt(lmeSt, lmePars) :
Calloc could not allocate (89359209 of 8) memory
In addition: Warning message:
Reached total allocation of 254Mb: see help(memory.size)
***********************************************************
My response variable is number of aphids/colony on shaded cocoa trees,
measured over 15 days (which I treat as a random factor) in four sites
(random factor), two levels of shade (fixed) within each site, and three
aphid predation treatmens (fixed) within each shade level. My total N =
2,256 observations (actually counts).
The comand lines were the same as in the example in:
Crawley, M. J. 2002. Statistical Computing: an intro to data analysis
using S-Plus. Wiley. pages: 702-704.
My questions are:
1-Did I use the appropriate analysis to my data?
2-Is the lack of memory caused by my large N and many factor interactions?
3-Since I have counts, should I specify (family=poisson) in the model? If
so, where in the command line this term must be located?
I am a complete beginner with R, so any help will be wellcome.
Evandro Silva
Assistant Professor
Universidade Estadual de Feira de Santana
Laboratory of Insect Ecology
Feira de Santana, BA, Brazil