Comparative analysis of classification methods in determining non-active student characteristics in Indonesia Open University

Dewi Juliah Ratnaningsih and Imas Sukaesih Sitanggang

Journal of Applied Statistics, 2016, vol. 43, issue 1, 87-97

Abstract: Classification is a data mining technique that aims to discover a model from training data that distinguishes records into appropriate classes. Classification methods can be applied in education, to classify non-active students in higher education programs based on their characteristics. This paper presents a comparison of three classification methods: Naïve Bayes, Bagging, and C4.5. The criteria used to evaluate performance of three classifiers are stratified cross-validation, confusion matrix, ROC curve, recall, precision, and F-measure. The data used for this paper are non-active students in Indonesia Open University (IOU) for the period of 2004--2012. The non-active students were divided into three groups: non-active students in the first three years, non-active students in first five years, and non-active students over five years. Results of the study show that the Bagging method provided a higher accuracy than Naïve Bayes and C4.5. The accuracy of bagging classification is 82.99%, while the Naïve Bayes and C4.5 are 80.04% and 82.74%, respectively. The classification tree resulted from the Bagging method has a large number of nodes, so it is quite difficult to use in decision-making. For that, the C4.5 tree is used to classify non-active students in IOU based in their characteristics.

Date: 2016
References: View complete reference list from CitEc
Citations:

Downloads: (external link)
http://hdl.handle.net/10.1080/02664763.2015.1077940 (text/html)
Access to full text is restricted to subscribers.

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:taf:japsta:v:43:y:2016:i:1:p:87-97

Ordering information: This journal article can be ordered from
http://www.tandfonline.com/pricing/journal/CJAS20

DOI: 10.1080/02664763.2015.1077940

Access Statistics for this article

Journal of Applied Statistics is currently edited by Robert Aykroyd

More articles in Journal of Applied Statistics from Taylor & Francis Journals
Bibliographic data for series maintained by Chris Longhurst ().