ML Anthology

Comparing Systems and Analyzing Functions to Improve Constructive Induction

Larry A. Rendell
ICML 1989 pp. 461-464
doi:10.1016/B978-1-55860-036-2.50117-X /icml/1989/rendell1989icml-comparing/

Abstract

Viewing a concept or any goal as a function over a description space suggests that a purpose of constructive induction is to transform that space into a simpler one by merging dispersed disjuncts, peaks, or regions of uniform value. Such a simplification of the goal function allows better task performance or improved behavior of associated learning. This is particularly relevant in empirical concept learning, where constructive induction is essential to learn hard concepts having many peaks. This analysis helps to unify current systems and suggests ways to improve their design.

ICML Semantic Scholar

Cite

Text

Rendell. "Comparing Systems and Analyzing Functions to Improve Constructive Induction." International Conference on Machine Learning, 1989. doi:10.1016/B978-1-55860-036-2.50117-X

Markdown

[Rendell. "Comparing Systems and Analyzing Functions to Improve Constructive Induction." International Conference on Machine Learning, 1989.](https://mlanthology.org/icml/1989/rendell1989icml-comparing/) doi:10.1016/B978-1-55860-036-2.50117-X

BibTeX

@inproceedings{rendell1989icml-comparing,
  title     = {{Comparing Systems and Analyzing Functions to Improve Constructive Induction}},
  author    = {Rendell, Larry A.},
  booktitle = {International Conference on Machine Learning},
  year      = {1989},
  pages     = {461-464},
  doi       = {10.1016/B978-1-55860-036-2.50117-X},
  url       = {https://mlanthology.org/icml/1989/rendell1989icml-comparing/}
}