Improved predictive models by indirect classification and bagging
for classification, regression and survival problems as well as
resampling based estimators of prediction error.
