Classication of Status of the Region on Java Island using C4.5, CHAID, and CART Methods

International Conference on Mathematics: Education, Theory and Application

数学;教育

Syaraswati, R.A.^1 ; Slamet, I.^2 ; Winarno, B.^1

Mathematics Department, Faculty of Mathematics and Natural Sciences, Universitas Sebelas Maret, Indonesia^1

Statistics Department, Faculty of Mathematics and Natural Sciences, Universitas Sebelas Maret, Indonesia^2

关键词: Binary decision trees;    Dependent variables;    Domestic product;    Economic growth rate;    Economic growths;    Economic success;    Manufacturing sector;    Nonparametric approaches;   

Others  :  https://iopscience.iop.org/article/10.1088/1742-6596/855/1/012053/pdf
DOI  :  10.1088/1742-6596/855/1/012053

学科分类：发展心理学和教育心理学

来源: IOP

【 摘 要 】

The indicator of region economic success can be measured by economic growth, presented by value of Gross Regional Domestic Product (GRDP). Java island has the biggest GDP contribution toward the Indonesian government, but not all of the region gives equality contribution. The C4.5, CHAID, and CART methods can be used for classifying the status of the region with nonparametric approach. The C4.5 and CHAID methods are non-binary decision tree, meanwhile the CART methods is binary decision tree. The purposes of this paper are to know how the classification and to determine the factors that influence on classification of the region. The dependent variable is status of the region which is divided into four categories based on Klassen typology. The result shows factors that have the biggest contribution on classification of status of the region on Java island based on C4.5 method are economic growth rate, electricity, gas, and water sector, and area. The factors that have the biggest contribution based on CHAID method are growth rate, manufacturing sector, and electricity, gas, and water sector, while based on CART method are growth rate, manufacturing sector, and electricity, gas, and water sector.