PRINCIPLED INDUCTION FROM FEATURE VALUES
Date
1992-07-01
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
Principled induction is defined in this thesis as the
process of arbitrarily selecting a valid description of a set
of examples, and gradually simplifying this description until
it is minimal in the preorder defined by the available simplification
operators. If the preorder is well chosen, principled induction
is a computationally feasible method for inducing descriptions.
Algorithms are presented that perform principled induction of
decision trees and rules from examples. These algorithms and
empirical results obtained using them are shown to support
the claim that principled induction is a clear and effective
representation of the problem of learning concept descriptions
from examples.
Description
Keywords
Computer Science