Alhakbani, Haya Abdullah and Al-Rifaie, Mohammad Majid ORCID: 0000-0002-1798-9615 (2016) A swarm intelligence approach in undersampling majority class. In: 10th International Conference, ANTS 2016, Brussels, Belgium, September 7-9, 2016, Proceedings. Lecture Notes in Computer Science, 9882 . Springer, Cham, Switzerland, pp. 225-232. ISBN 978-3319444277 ISSN 0302-9743 (Print), 1611-3349 (Online) (doi:https://doi.org/10.1007/978-3-319-44427-7_19)

Preview

PDF (Author's Accepted Manuscript) 20995 AL-RIFAIE_Swarm_Intelligence_Approach_In_Undersampling_Majority_Class_(AAM)_2016.pdf - Accepted Version Download (364kB) | Preview

Abstract

Over the years, machine learning has been facing the issue of imbalance dataset. It occurs when the number of instances in one class significantly outnumbers the instances in the other class. This study investigates a new approach for balancing the dataset using a swarm intelligence technique, Stochastic Diffusion Search (SDS), to undersample the majority class on a direct marketing dataset. The outcome of the novel application of this swarm intelligence algorithm demonstrates promising results which encourage the possibility of undersampling a majority class by removing redundant data whist protecting the useful data in the dataset. This paper details the behaviour of the proposed algorithm in dealing with this problem and investigates the results which are contrasted against other techniques.

Item Type:	Conference Proceedings
Title of Proceedings:	10th International Conference, ANTS 2016, Brussels, Belgium, September 7-9, 2016, Proceedings
Uncontrolled Keywords:	swarm intelligence, class imbalance, stochastic diffusion search, SVM
Subjects:	Q Science > QA Mathematics > QA75 Electronic computers. Computer science
Faculty / School / Research Centre / Research Group:	Faculty of Liberal Arts & Sciences > Computational Science & Engineering Group (CSEH) Faculty of Engineering & Science > School of Computing & Mathematical Sciences (CMS) Faculty of Engineering & Science
Last Modified:	04 Mar 2022 13:07
URI:	http://gala.gre.ac.uk/id/eprint/20995

Actions (login required)

View Item

Downloads