forum.alglib.net

I am training a random forest with 10 features and 1 continuous target variable. When I train the random forest with 10 trees, everything works successfully. When I increase the number of trees to 50 or 100, I get ap_error.

my code is:

ae_int_t ntrees = 50;
ae_int_t mtry = 6;
fraction = 0.5;
decisionforestbuilder builder;
ae_int_t nvars = 10;
ae_int_t nclasses = 1;
ae_int_t npoints = train_size;

dfbuildercreate(builder);
dfbuildersetdataset(builder, train_data, npoints, nvars, nclasses);
dfbuildersetrndvars(builder,mtry);
dfbuildersetseed(builder, 1);
dfbuildersetsubsampleratio(builder, fraction);
dfreport rep;
dfbuilderbuildrandomforest(builder, ntrees, model, rep);

What can this error be?

Using a try and catch block, I got:


error msg: ae_malloc(): out of memory

What should I do?