The current randomForest package stores the entire proximity matrix (30000 x 30000 in your case), which is needed for imputation. Breiman and Cutler's Fortran code stores the largest nrnn element of each row. If you really want to do it that way, use the Fortran version.
Andy From: Zhu Ailing > > HI, > I have data set of 30000 record with 162 features. when I > try to fill out the missing values using rfImpute(), I got > kicked out becz: > tra.imputed <- rfImpute(tra.na[,-163],tra.na[,163],iter=5,ntree=10) > Error in matrix(0, n, n) : cannot allocate vector of length 900000000 > > I wonder how to set the parameters(nrnn) for computing the > PROXIMITY. thanks iris > > -----Original Message----- > From: Zhu Ailing > Sent: Wed 3/29/2006 11:41 AM > To: '[EMAIL PROTECTED]' > Cc: > Subject: missing value replacement for test data in > random forest > > > Hi, > > In R, how to do missing value replacement for test data > in randome forest in the way Breiman decribed. > > thanks in advance > > iris > > ______________________________________________ > [email protected] mailing list > https://stat.ethz.ch/mailman/listinfo/r-help > PLEASE do read the posting guide! > http://www.R-project.org/posting-guide.html > > ______________________________________________ [email protected] mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
