What steps will reproduce the problem?
1. attempt to train a RF with a high-dimensional dataset (34
1300-dimensional vectors), using 101 trees and mtry=200 features:
myRF = classRF_train(foo_vecs(2:35,1:1300),foo_classLabels(2:35,:),101,200);
foo_vecs is a 36 x 4005 matrix of doubles
foo_classLabels is a 36 x 1 vector of doubles (-1,+1)
(see attached file)
What is the expected output? What do you see instead?
Expected: a trained RF.
Instead: a segmentation violation, with stack trace:
[0] mexClassRF_train.mexmaci64:makeA(double*, int, int, int*, int*,
int*)~ + 151 bytes
[1] mexClassRF_train.mexmaci64:classRF(double*, int*, int*, int*, int*,
int*, int*, int*, int*, int*, int*, int*, double*, double*, int*, int*,
int*, double*, double*, double*, double*, int*, int*, int*, int*, int*,
int*, double*, double*, int*, double*, int*, int*, double*, int*, int,
double*, double*, int*)~ + 2673 bytes
[2] mexClassRF_train.mexmaci64:mexFunction~ + 3192 bytes
... more stuff
What version of the product are you using? On what operating system?
Version svn-v8? (0.02), MacOSX 10.6.2, Matlab 7.9.0.529 (R2009b) 64 bits,
Intel Core 2 Duo (x86 Family 6 Model 7 Stepping 10). Mex file compiled from
source.
Note: the mex file works fine until the training set reaches about 34 x
1200, thereafter crashes. Could it be a memory allocation issue?