Reference Hub1
Variant of Northern Bald Ibis Algorithm for Unmasking Outliers

Variant of Northern Bald Ibis Algorithm for Unmasking Outliers

Ravi Kumar Saidala
Copyright: © 2020 |Volume: 12 |Issue: 1 |Pages: 15
ISSN: 1942-9045|EISSN: 1942-9037|EISBN13: 9781799806097|DOI: 10.4018/IJSSCI.2020010102
Cite Article Cite Article

MLA

Saidala, Ravi Kumar. "Variant of Northern Bald Ibis Algorithm for Unmasking Outliers." IJSSCI vol.12, no.1 2020: pp.15-29. http://doi.org/10.4018/IJSSCI.2020010102

APA

Saidala, R. K. (2020). Variant of Northern Bald Ibis Algorithm for Unmasking Outliers. International Journal of Software Science and Computational Intelligence (IJSSCI), 12(1), 15-29. http://doi.org/10.4018/IJSSCI.2020010102

Chicago

Saidala, Ravi Kumar. "Variant of Northern Bald Ibis Algorithm for Unmasking Outliers," International Journal of Software Science and Computational Intelligence (IJSSCI) 12, no.1: 15-29. http://doi.org/10.4018/IJSSCI.2020010102

Export Reference

Mendeley
Favorite Full-Issue Download

Abstract

Clustering, one of the most attractive data analysis concepts in data mining, are frequently used by many researchers for analysing data of variety of real-world applications. It is stated in the literature that traditional clustering methods are trapped in local optima and fail to obtain optimal clusters. This research work gives the design and development of an advanced optimum clustering method for unmasking abnormal entries in the clinical dataset. The basis is the NOA, a recently proposed algorithm, driven by mimicking the migration pattern of Northern Bald Ibis (Threskiornithidae) birds. First, we developed the variant of the standard NOA by replacing C1 and C2 parameters of NOA with chaotic maps turning it into the VNOA. Later, we utilized the VNOA in the design of a new and advanced clustering method. VNOA is first benchmarked on a 7 unimodal (F1–F7) and 6 multimodal (F8–F13) mathematical functions. We tested the numerical complexity of proposed VNOA-based clustering methods on a clinical dataset. We then compared the obtained graphical and statistical results with well-known algorithms. The superiority of the presented clustering method is evidenced from the simulations and comparisons.

Request Access

You do not own this content. Please login to recommend this title to your institution's librarian or purchase it from the IGI Global bookstore.