The advantage of the out-of-bag error is that it is a more realistic estimate of the error rate that is to be expected in a new test sample, than the naive and over-optimistic estimate of the error rate resulting from the prediction of the entire learning sample (Breiman 1996b) (see also Boulesteix, Strobl, Augustin, and Daumer (2008) for a review on resampling-based error estimation). The standard and out-of-bag prediction accuracy of a random forests with ntree=500 and mtry=2 for our smoking data example is 74.5% and 71.5% respectively, where the out-of-bag prediction accuracy is more conservative.
