BAGGING PREDICTORS

Bagging predictors is a method for generating multiple versions of a p redictor and using these to gel an aggregated predictor. The aggregati on averages over the versions when predicting a numerical outcome and does a plurality vote when predicting a class. The multiple versions a re formed by making bootstrap replicates of the learning set and using these as new learning sets. Tests on real and simulated data sets usi ng classification and regression trees and subset selection in linear regression show that bagging can give substantial gains in accuracy. T he vital element is the instability of the prediction method. If pertu rbing the learning set can cause significant changes in the predictor constructed, then bagging can improve accuracy.