A methodology for first order learning : a case study

Wim Van Laer and Luc De Raedt. In F. Verdenius and W. van den Broek, editors, Eighth Belgian-Dutch Conference on Machine Learning, pages 105--113. ATO-DLO, November 1998.

Abstract

Many well-known learning systems like \CNT\ and C4.5 use an attribute-value representation for the examples. In Inductive Logic Programming (ILP) a first-order logical formalism is used, which is a more expressive representation. We will introduce a 5-step methodology to upgrade an existing propositional learner towards a first-order learning system. The system \Icl\ (\em Inductive Classification Logic) is used as a case study.

BibTeX entry.

Other publications

by Wim Van Laer,
by L De Raedt,
on ILP Theory,
of 1998.

Wim Van Laer, Wim.VanLaer@cs.kuleuven.ac.be,