A case study on machine learning and classification

Amit Kumar and Bikash Kanti Sarkar

International Journal of Information and Decision Sciences, 2017, vol. 9, issue 2, 179-208

Abstract: As a young research field, the machine learning has made significant progress and covered a broad spectrum of applications for the last few decades. Classification is an important task of machine learning. Today, the task is used in a vast array of areas. The present article provides a case study on various classification algorithms (under machine learning), their applicability and issues. More specifically, a step by step progress on this area is discussed in this paper. Further, an experiment is conducted over 12 real-world datasets drawn from University of California, Irvine (UCI, a machine learning repository) using four competent individual learners namely, C4.5 (decision tree-based classifier), Naïve Bayes, k-nearest neighbours (k-NN), neural network and two hybrid learners: Bagging (based on decision tree) and (fuzzy + rough-set + k-NN: a hybrid system) for head to head comparison of their classification performance. Their merits and demerits (as discussed in this article) are analysed accordingly with the obtained results.

Keywords: machine learning; classification; applicability; performance; hybrid; prediction; accuracy; classifier. (search for similar items in EconPapers)
Date: 2017
References: Add references at CitEc
Citations: View citations in EconPapers (1)

Downloads: (external link)
http://www.inderscience.com/link.php?id=84885 (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:ids:ijidsc:v:9:y:2017:i:2:p:179-208

Access Statistics for this article

More articles in International Journal of Information and Decision Sciences from Inderscience Enterprises Ltd
Bibliographic data for series maintained by Sarah Parker ().