Abstract

Bootstrapped aggregation of classifiers, also referred to as bagging, is a classic meta-classification algorithm. We extend it to a two-stage architecture consisting of an initial voting amongst one-versus-all classifiers or single-class recognizers, and a second stage of one-versus-one classifiers or two-class discriminators used for disambiguation. Since our method constructs an ensemble of elementary classifiers, it lends itself very well to parallelization. We describe a static workload balancing strategy for embarrassingly parallel classifier construction as well as a parallelization of the classification process with the message passing interface. In our experiments, which are evaluated in terms of classification performance and speed-up, we obtained an up to three-fold increase in precision and significantly increased recall values.

Keywords

Cite this article

@article{cai-842,
  author  = {Verena Christina Horak and Tobias Berka and Marian Vajteršic},
  title   = {Parallel Classification with Two-Stage Bagging Classifiers},
  journal = {Computing and Informatics},
  volume  = {32},
  number  = {4},
  pages   = {661--677},
  year    = {2013},
  issn    = {1335-9150},
  url     = {https://cai.type.sk/content/2013/4/parallel-classification-with-two-stage-bagging-classifiers/}
}
Download .bib