> Ot should be noted that the performance of single split into training + test > does not perform satisfactorily unless N > 10,000 in many cases.
Agreed. I've had a few cases where the resampled statistics on the training set look poor but the test set results are much better. In the end, the training set statistics were more indicative of future performance. The train() function in caret has a few options for resampling models... -- Max ______________________________________________ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
