DocumentCode :
798534
Title :
Class-dependent discretization for inductive learning from continuous and mixed-mode data
Author :
Ching, John Y. ; Wong, Andrew K.C. ; Chan, Keith C C
Author_Institution :
Dept. of Syst. Design Eng., Waterloo Univ., Ont., Canada
Volume :
17
Issue :
7
fYear :
1995
fDate :
7/1/1995 12:00:00 AM
Firstpage :
641
Lastpage :
651
Abstract :
Inductive learning systems can be effectively used to acquire classification knowledge from examples. Many existing symbolic learning algorithms can be applied in domains with continuous attributes when integrated with a discretization algorithm to transform the continuous attributes into ordered discrete ones. In this paper, a new information theoretic discretization method optimized for supervised learning is proposed and described. This approach seeks to maximize the mutual dependence as measured by the interdependence redundancy between the discrete intervals and the class labels, and can automatically determine the most preferred number of intervals for an inductive learning application. The method has been tested in a number of inductive learning examples to show that the class-dependent discretizer can significantly improve the classification performance of many existing learning algorithms in domains containing numeric attributes
Keywords :
learning by example; maximum entropy methods; pattern classification; probability; class-dependent discretization; classification knowledge; continuous data; inductive learning; information theoretic discretization method; interdependence redundancy; mixed-mode data; numeric attributes; supervised learning; Discrete transforms; Entropy; Learning systems; Machine learning; Machine learning algorithms; Mutual information; Optimization methods; Performance evaluation; Supervised learning; Testing;
fLanguage :
English
Journal_Title :
Pattern Analysis and Machine Intelligence, IEEE Transactions on
Publisher :
ieee
ISSN :
0162-8828
Type :
jour
DOI :
10.1109/34.391407
Filename :
391407
Link To Document :
بازگشت