Collective Agreement-based Pruning of EnsemblesCollective Agreement-Based Pruning of Ensembles • October 3rd, 2009
Contract Type FiledOctober 3rd, 2009The main idea of ensemble methodology is to weigh several individual pattern classifiers, and combine them to reach a better classification performance. Nevertheless, some ensembles superfluously contain too many members, which results in large storage requirements and in some cases it may even reduce classification performance. The goal of ensemble pruning is to identify a subset of ensemble members that performs at least as good as the original ensemble and discard any other members as redundant members.