[BioC] can't allocate vector for random forest

Martin Morgan mtmorgan at fhcrc.org
Sun Mar 4 17:57:30 CET 2012


On 03/04/2012 07:49 AM, Alpesh Querer wrote:
> from the R side, try increasing the memory limit.
> You can check default memory by the command memory.limit()
> Alternatively, you can change the memory limit within R using the
> memory.limit function, giving a new limit in Mb. (For example, type
> memory.limit(2048) to change the memory limit to 2 Gb.)

It's useful to remember that this kind of error message often means 
"I've already allocated a lot of memory, and cannot allocate an 
additional 1.8G". It is often useful to filter the input to machine 
learning algorithms, e.g., selecting a subset of the most variable (and 
hence most statistically informative, by a robust measure of 
variability) genes. See the genefilter package.

Martin

>
>
> On Sun, Mar 4, 2012 at 12:05 AM, Salwa Eid<salwaeid at hotmail.com>  wrote:
>
>>
>>
>> limit
>>
>> Dear all,    Have read dna microarray cel files and normalized the data
>> and then tried to use the random forest classifier but i keep getting the
>> error "cannot allocate vector of 1.8G".  I had a 3gb ram and upgraded it to
>> 8gb and am still getting the same error exactly.  In the first case got a
>> warning that it reached total allocation of 3gb and the second case got the
>> warning, reached total allocation of 8gb.   The code I used for random
>> forest is test<-randomForest(Type~., data=tdata) The tdata is 22215 genes
>> for four samples and the type of each sample(2 classes).  Prediction here
>> is done for the type.  The tdata is a matrix of 22216 columns which
>> consists of 22215 genes and the type and there are 4 rows for four samples.
>>   Two types. Any help? Regards,salwa
>>         [[alternative HTML version deleted]]
>>
>> _______________________________________________
>> Bioconductor mailing list
>> Bioconductor at r-project.org
>> https://stat.ethz.ch/mailman/listinfo/bioconductor
>> Search the archives:
>> http://news.gmane.org/gmane.science.biology.informatics.conductor
>>
>
> 	[[alternative HTML version deleted]]
>
> _______________________________________________
> Bioconductor mailing list
> Bioconductor at r-project.org
> https://stat.ethz.ch/mailman/listinfo/bioconductor
> Search the archives: http://news.gmane.org/gmane.science.biology.informatics.conductor


-- 
Computational Biology
Fred Hutchinson Cancer Research Center
1100 Fairview Ave. N. PO Box 19024 Seattle, WA 98109

Location: M1-B861
Telephone: 206 667-2793



More information about the Bioconductor mailing list