Re: [R] batch job GLM calculations

From: Milton Cezar Ribeiro <>
Date: Mon 15 Jan 2007 - 12:26:16 GMT


  Someone help-me on this task some weeks ago, and it is working find to me. I use something like:

  findmodels <- function(modeltype = "lm", dataset, pattern) {  ls <- ls(".GlobalEnv", pattern=pattern)  mods <- ls[sapply(ls, function(x) inherits(get(x), modeltype))]  if (!missing(dataset)) { <- function(x) as.character(x$call[["data"]])  mods <- mods[sapply(mods, function(x) == dataset)]

 models <- lapply(mods, get)
 class(models) <- c("ensemble", class(models))  models


models<- findmodels(pattern="mod_")

  So you can use a "for (i in 1:length(models))" to extract the statistics tha you want for each model and save it in a file (I´m saving it easily in Access using RODBC).    

  I hope this help,    


Indermaur Lukas <> escreveu:   Hello

I want to batch job the calculation of many GLM-models, extract some values and store them in a file. Almost everything in the script below works (read file, extract values and write them to file) except I fail in indexing the GLM with the modelstructure it should run. Running GLM's conventionally is no problem.

Conventionally a GLM is calculated as:

glm(ZlogHRS ~ ZRi+ZE+ZPROX_MN+ZED+ZAlwd+ZT2+ZW+ZN+Sex+y, family = gaussian,

(just a note: dependent variable is ZlogHRS, while the others are indepent variables)

Desired way: sequentially run GLM

I want R to take the model structure to take from a vector called "modelstructure" and paste it into the GLM like:

glm(modelstructure[i], family = gaussian,

It would considerably ease my workload if there is a solution to the indexing problem within a GLM. I appreciate any hint.

Best regards



my R skills are rather poor

----------------------------------------START R-CODE-----------------------------------

# Read file

t.url <- "C://HR_calculations/2005_2006/HR_SIZE/Kandidatenmodelle_Berechnung/inputfiles/"

t.tuti <- read.table(paste(t.url, "All_animals.txt", sep=""),header=T)

collect.results <- function(x) {

#resets vectors which will be filled

i <- 0


#put models names hierarchically in vector

modelnames <- c("1=global", "2=biotic1", "3=biotic2", "4=abiotic") #keep track of changes in model names and number

for (i in 1:length(modelnames))

#model structure of the four models given for all models to run


modelstructure <- c(

"ZlogHRS ~ ZRi+ZE+ZPROX_MN+ZED+ZAlwd+ZT2+ZW+ZN+Sex+y",


"ZlogHRS ~ ZRi",




"ZlogHRS ~ ZE")

ts.model <- glm(modelstructure[i], family = gaussian,

# Extracts some mode results

n[i] <- length(resid(ts.model))

AICA[i] <- AIC(ts.model)


#Writes results to data file

x = data.frame(

modelnames, n, AICA)

write.table(x, paste(t.url, file="Results.txt"), sep="\t", quote=F)

----------------------------------------END R-CODE-----------------------------------

Lukas Indermaur, PhD student
eawag / Swiss Federal Institute of Aquatic Science and Technology ECO - Department of Aquatic Ecology
Überlandstrasse 133
CH-8600 Dübendorf

Phone: +41 (0) 71 220 38 25
Fax : +41 (0) 44 823 53 15
Email: mailing list PLEASE do read the posting guide and provide commented, minimal, self-contained, reproducible code.

        [[alternative HTML version deleted]] mailing list PLEASE do read the posting guide and provide commented, minimal, self-contained, reproducible code. Received on Mon Jan 15 23:32:47 2007

Archive maintained by Robert King, hosted by the discipline of statistics at the University of Newcastle, Australia.
Archive generated by hypermail 2.1.8, at Mon 15 Jan 2007 - 13:30:32 GMT.

Mailing list information is available at Please read the posting guide before posting to the list.