Hello,
In predict.glm, there's no argument 'data', it's
'newdata'.
As for your problem, maybe try doing one prediction at a time, write the
results to file, then the next.
Hope this helps,
Rui Barradas
Em 22-05-2013 15:20, Kristi Glover escreveu:> Hi R user,
> I was trying to develop a model (logistic regression) for 4001 dependent
variables using 15 environmental variables (45000 rows); and then trying to use
the models to predict in future. I used following code but it took so much time
and consumed 100% of the PC memory. Even though- analysis was not complete. I
got a following message
> " Reached total allocation of 8098Mb: see help(memory.size)". I
increased memory size to 8GB. but still I could not complete the analysis.
> Any suggestion to reduce the memory and compute the big data set.
>
> #------------------------------------------------------------------
> data=spec.Env
>
> models <- list()
> PredictModelsCur<-list()
> PredictModelsA1<-list()
> PredictModelsA2<-list()
> PredictModelsA3<-list()
> dvnames <- paste("V", 2:4003, sep="")
> ivnames <- paste("env", 1:15,
sep="",collapse="+") ## for some value of n
>
> for (y in dvnames){
> form <- formula(paste(y,"~",ivnames))
> models[[y]] <- glm(form, data=spec.Env, family='binomial')
> PredictModelsCur[[y]]<-predict(models[[y]],
type="response")
> PredictModelsA1[[y]]<-predict(models[[y]], data = a1.Futute,
type="response")
> PredictModelsA2[[y]]<-predict(models[[y]], data = a2.Futute,
type="response")
> PredictModelsA3[[y]]<-predict(models[[y]], data = a3.Futute,
type="response")
> }
>
> write.csv(PredictModelsCur, "PredictModelsCur.csv", row.names=F)
> write.csv(PredictModelsA1, "PredictModelsA1.csv", row.names=F)
> write.csv(PredictModelsA2, "PredictModelsA2.csv", row.names=F)
> write.csv(PredictModelsA3, "PredictModelsA3.csv", row.names=F)
>
>
> [[alternative HTML version deleted]]
>
> ______________________________________________
> R-help at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide
http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.
>