A Discretization Algorithm Based on Gini Criterion

Author

Zhang, Xiao-hang ; Wu, Jun ; Lu, Ting-jie ; Jiang, Yuan

Author_Institution

Beijing Univ. of Posts & Telecommun., Beijing

Volume

5

fYear

2007

fDate

19-22 Aug. 2007

Firstpage

2557

Lastpage

2561

Abstract

In this paper, a supervised, global and static algorithm for the discretization of continuous attributes is presented. This algorithm takes account of the distribution of class probability vector by applying the Gini criterion. The proposed discretization method is compared with Ent-MDLP, which is known as one of the best discretization methods, in terms of predictive error rate and tree size. This paper reveals that the proposed algorithm is effective and can be a good alternative to the entropy-based discretization methods in some situations.

Keywords

learning (artificial intelligence); probability; Gini criteria; class probability vector; discretization algorithm; global algorithm; predictive error rate; static algorithm; supervised algorithm; tree size; Conference management; Cybernetics; Data preprocessing; Economic forecasting; Error analysis; Frequency estimation; Machine learning; Machine learning algorithms; Partitioning algorithms; Spatial databases; Data preprocessing; Discretization; Gini criterion;

fLanguage

English

Publisher

ieee

Conference_Titel

Machine Learning and Cybernetics, 2007 International Conference on

Conference_Location

Hong Kong

Print_ISBN

978-1-4244-0973-0

Electronic_ISBN

978-1-4244-0973-0

Type

conf

DOI

10.1109/ICMLC.2007.4370578

Filename

4370578