The following information was submitted:
Transactions: INTERNATIONAL JOURNAL of MATHEMATICAL MODELS AND METHODS IN APPLIED SCIENCES
Transactions ID Number: 20-582
Full Name: Hyontai Sug
Position: Associate Professor
Age: ON
Sex: Male
Address: Division of Computer and Information Engineering, Dongseo University, Busan, 617-716
Country: KOREA
Tel: 82-51-320-1733
Tel prefix:
Fax: 82-51-327-8955
E-mail address: hyontai@yahoo.com
Other E-mails: shtdaum@hanmail.net
Title of the Paper: Towards More Accurate Classification of Instances in Minor Classes
Authors as they appear in the Paper: hyontai sug
Email addresses of all the authors: hyontai@yahoo.com
Number of paper pages: 8
Abstract: In the task of data mining using decision trees, the classification accuracy for minor classes is usually poorer than that of major classes, because decision trees are built to optimize accuracy throughout the available data set and the number of instances belonging to minor classes is relatively rare. So the instances in minor classes are treated less importantly in classification. This paper suggests a method based on progressive over-sampling with respect to minor classes to generate more accurate decision trees for the minor classes for the case that we need more accurate classification for the minor classes. Experiments were done with two representative decision tree algorithms, C4.5 and CART, and two data sets, ¡®adult¡¯ and ¡®internet ads¡¯, and showed the validity of the method.
Keywords: Biased sampling, minor classes, data mining.
EXTENSION of the file: .pdf
Special (Invited) Session: Improving the Performance of Minor Class in Decision Tree Using Duplicating Instances
Organizer of the Session: 650-628
How Did you learn about congress:
IP ADDRESS: 118.39.140.117
