Hello, I have a rather large set of data I need to analyze, currently I need to
work with a 200000 by 200000 matrix, I'm using the package bigmemory but so
far I can only allocate a 66000 by 66000 matrix, when I increase those values I
get the following error:
> AdjMat <- big.matrix(nrow=68000,ncol=68000)
Cannot allocate memory
BigMatrix.cpp line 225
Error in big.matrix(nrow = 68000, ncol = 68000) :
Error: memory could not be allocated for instance of type big.matrix
As a part of my analyisis I need to calculate de correlation coefficient, but
when I try to do that in a "smaller" matrix I get this other error.
> A <- big.matrix(nrow=45000,ncol=45000)
> AM <- (cor(A, method="pearson"))
Error in as.vector(data) :
no method for coercing this S4 class to a vector
Calls: cor ... as.matrix -> as.matrix.default -> array -> as.vector
In addition: Warning message:
In is.na(x) : is.na() applied to non-(list or vector) of type 'S4'
Execution halted
My question is, how big of a matrix can I have when I use the package bigmemory,
and if I use such a package, can I use the function cor()? If I can't, then
do you have any suggestions on how can I address this problem?
Thank you very much.
ERV
[[alternative HTML version deleted]]
