A Bipolar Interpretation of Fuzzy Decision Trees - Data Mining: Foundations and Practice - page 108

A Bipolar Interpretation of Fuzzy Decision
Trees ∗

Tuan-Fang Fan 1 , Churn-Jung Liau 2 , and Duen-Ren Liu 1

1

Institute of Information Management, National Chiao-Tung University,
Hsinchu 300, Taiwan
[e-mail], [e-mail]

2

Institute of Information Science, Academia Sinica, Taipei 115, Taiwan
[e-mail]

Summary. Decision tree construction is a popular approach in data mining and
machine learning, and some variants of decision tree algorithms have been proposed
to deal with different types of data. In this paper, we present a bipolar interpretation
of fuzzy decision trees. With the interpretation, various types of decision trees can
be represented in a unified form. The edges of a fuzzy decision tree are labeled by
fuzzy decision logic formulas and the nodes are split according to the satisfaction of
these formulas in the data records. We present a construction algorithm for general
fuzzy decision trees and show its application to different types of training data.

1 Introduction

Decision trees play an important role in machine learning and data mining
[11, 13-15], and classifiers based on decision trees work well for precise data.
However, imprecision, uncertainty, and incompleteness prevail in real-world
data. To deal with different kinds of uncertainty, a variety of modifications
and generalizations of decision trees, such as the fuzzy decision tree [6] and
the multi-valued decision tree [2], have been proposed.
As these variants of decision trees were proposed independently, there has
not been a comparative study using a common framework. In this paper, we
try to address this situation by proposing a bipolar interpretation of fuzzy
decision trees. The interpretation makes it possible to represent various types
of decision trees in a uniform framework. Our framework can deal with very
general forms of fuzzy data and fuzzy rules. When specialized for specific

∗ Preliminary results of this paper were presented at the IEEE 2003 ICDM workshop
on Foundations and New Directions in Data Mining and appeared in the informal
proceedings of the workshop with limited distribution.
T.-F. Fan et al.: A Bipolar Interpretation of Fuzzy Decision Trees , Studies in Computational