Knowledge reuse in multiple classifier systems 1 Kurt

We introduce a framework for the reuse of knowledge from previously trained classifiers to improve performance in a current and possibly related classification task. The approach used is flexible in the type and relevance of reused classifiers and is also scalable. Experiments on public domain data sets demonstrate the usefulness of this approach when one is faced with very few training examples or very noisy training data. q 1997 Published by Elsevier Science B.V.