ML Anthology

Informed Pruning in Constructive Induction

George Drastal
ICML 1991 pp. 132-136
doi:10.1016/B978-1-55860-200-7.50030-1 /icml/1991/drastal1991icml-informed/

Abstract

This paper concerns the effect of training noise on performance in a constructive induction algorithm. We give a method for using domain theory bias in postprocessing a classification rule to correct for overfltting. The algorithm is a variant of the nearest-neighbor algorithm used in digital communication to decode a block-coded message. Experimental results are presented.

ICML Semantic Scholar

Cite

Text

Drastal. "Informed Pruning in Constructive Induction." International Conference on Machine Learning, 1991. doi:10.1016/B978-1-55860-200-7.50030-1

Markdown

[Drastal. "Informed Pruning in Constructive Induction." International Conference on Machine Learning, 1991.](https://mlanthology.org/icml/1991/drastal1991icml-informed/) doi:10.1016/B978-1-55860-200-7.50030-1

BibTeX

@inproceedings{drastal1991icml-informed,
  title     = {{Informed Pruning in Constructive Induction}},
  author    = {Drastal, George},
  booktitle = {International Conference on Machine Learning},
  year      = {1991},
  pages     = {132-136},
  doi       = {10.1016/B978-1-55860-200-7.50030-1},
  url       = {https://mlanthology.org/icml/1991/drastal1991icml-informed/}
}