데이터 마이닝을 위한 이산화 알고리즘에 대한 비교 연구 = A Comparative Study on Discretization Algorithms for Data Mining

다국어 초록 (Multilingual Abstract)

The discretization process that converts continuous attributes into discrete ones is a preprocessing step in data mining such as classification. Some classification algorithms can handle only discrete attributes. The purpose of discretization is to obtain discretized data without losing the information for the original data and to obtain a high predictive accuracy when discretized data are used in classification. Many discretization algorithms have been developed. This paper presents the results of our comparative study on recently proposed representative discretization algorithms from the view point of splitting versus merging and supervised versus unsupervised. We implemented R codes for discretization algorithms and made them available for public users.

참고문헌 (Reference)

1 Ziarko,W, "Variable precision rough set model" 46 : 39-59, 1993

2 Jin, H, "Using AUC and accuracy in evaluating learning algorithms" 17 : 299-310, 2005

3 Merz, C. J, "UCI repository of machine learning database, department ofinformation and computer science"

4 Dougherty, J., "Supervised and unsupervised discretization of continuousfeatures"

5 Pawlak,Z, "Rough sets" 11 : 341-356, 1982

6 R Development Core Team, "R: A language and environment for statistical computing, R Foundationfor statistical computing, Vienna, Austria, ISBN 3-900051-07-0"

7 Fayyad, U. M, "Multi-interval discretization of continuous-valued attributes forclassification learning" 13 : 1022-1027, 1993

8 Tay, F. E. H, "Modified Chi2 algorithm for discretization" 14 : 666-670, 2002

9 Chmielewski, M. R, "Global discretization of continuous attributes aspreprocessing for machine learning" 15 : 319-331, 1996

10 Liu, H, "Feature selection and discretization" 9 : 642-645, 1997

11 Acuna,E, "Dprep: Data preprocessing and visualization functions for classification, R packageversion 1.0"

12 Kim, H. J, "Discretization: Data preprocessing, discretization for classification. R package version1.0"

13 Sotiris, K, "Discretization techiniques: A recent survey" 32 : 47-58, 2006

14 Liu, H, "Discretization : An enabling technique" 6 : 393-423, 2002

15 Witten, I. H, "Data Mining Practical Machine learning Tools and Techniques, Morgankaufmann"

16 Zhaoa, Y. H, "Comparison of decision tree methods for finding active objects" 41 : 1955-1959, 2008

17 Kerber,R, "ChiMerge: Discretization of numeric attributes" 123-128, 1992

18 Liu, H, "Chi2: Feature selection and discretization of numeric attributes"

19 Kurgan, L. A, "CAIM discretization algorithm" 16 : 145-153, 2004

20 Quinlan,R, "C4.5: Programs for Machine Learning" Morgan Kaufmann Publishers Inc 1993

21 Su, C. T, "An extended Chi2 algorithm for discretization of real value attributes" 17 : 437-441, 2005

22 Gonzalez-Abril, L, "Ameva: An autonomous discretizationalgorithm" 36 : 5327-5332, 2009

23 Ling, C. X., "AUC : A better measure than accuracy in comparing learningalgorithm" 2671 : 991-, 2003

24 Tsai, C. J, "A discretization algorithm based on class-attribute contingencycoefficient" 178 : 714-731, 2008