ML Anthology

NFDT: A System That Learns Flexible Concepts Based on Decision Trees for Numerical Attributes

Thierry Van de Merckt
ICML 1992 pp. 322-331
doi:10.1016/B978-1-55860-247-2.50046-2 /icml/1992/demerckt1992icml-nfdt/

Abstract

Most research efforts in Empirical Concept Learning have been devoted to nominal attribute spaces. Using numerical ordered spaces rises some specific problems, especially for Top-Down algorithms. A new system, called NFDT (Numerical Flexible Decision Tree), which aims to solve these problems, is presented in this paper. It builds a Flexible Matching function based on a Decision Tree issued from a classical Top-Down Induction algorithm. Empirical tests show that NFDT performs better than a Classical Decision Tree method, especially when noise is present, and that it can be used to evaluate the quality of concept descriptions issued from a Decision Tree.

ICML Semantic Scholar

Cite

Text

Van de Merckt. "NFDT: A System That Learns Flexible Concepts Based on Decision Trees for Numerical Attributes." International Conference on Machine Learning, 1992. doi:10.1016/B978-1-55860-247-2.50046-2

Markdown

[Van de Merckt. "NFDT: A System That Learns Flexible Concepts Based on Decision Trees for Numerical Attributes." International Conference on Machine Learning, 1992.](https://mlanthology.org/icml/1992/demerckt1992icml-nfdt/) doi:10.1016/B978-1-55860-247-2.50046-2

BibTeX

@inproceedings{demerckt1992icml-nfdt,
  title     = {{NFDT: A System That Learns Flexible Concepts Based on Decision Trees for Numerical Attributes}},
  author    = {Van de Merckt, Thierry},
  booktitle = {International Conference on Machine Learning},
  year      = {1992},
  pages     = {322-331},
  doi       = {10.1016/B978-1-55860-247-2.50046-2},
  url       = {https://mlanthology.org/icml/1992/demerckt1992icml-nfdt/}
}