A new enhancement of the k-NN algorithm by Using an optimization technique

Altayef, Ehsan, Anayi, Fatih ORCID: https://orcid.org/0000-0001-8408-7673 and Packianather, Michael ORCID: https://orcid.org/0000-0002-9436-8206 2022. A new enhancement of the k-NN algorithm by Using an optimization technique. Presented at: 2nd International Conference on Advance Computing and Innovative Technologies in Engineering (ICACITE), Greater Noida, India, 28-29 April 2022. Proceedings of 2nd International Conference on Advance Computing and Innovative Technologies in Engineering. IEEE, pp. 24-31. 10.1109/ICACITE53722.2022.9823537

Preview

PDF - Accepted Post-Print Version Download (735kB) | Preview

Abstract

Of a number of ML (Machine Learning) algorithms, k-nearest neighbour (KNN) is among the most common for data classification research, and classifying diseases and faults, which is essential due to frequent alterations in the training dataset, in which it would be expensive using most methods to construct a different classifier every time this happens. Therefore, KNN can be used effectively as it does not require a residual classifier to be constructed in advance. KNN offers ease of use and can be applied across a broad variation spectrum. Here, a novel KNN classification approach is put forward using the Bayesian Optimization Algorithm (BOA) for optimisation. This paper seeks to make classification more accurate and suggest alterations of nearest neighbour K value to use information about dataset structure and the similarity measure of distance. The findings of experimental work based on the University of California Irvine (UCI) repository datasets in general shows improved performance of classifiers compared with conventional KNN and give greater reliability without a significant time cost to speed.

Item Type:	Conference or Workshop Item (Paper)
Date Type:	Published Online
Status:	Published
Schools:	Engineering
Publisher:	IEEE
ISBN:	9781665437905
Date of First Compliant Deposit:	7 September 2022
Date of Acceptance:	25 January 2022
Last Modified:	28 Dec 2024 15:46
URI:	https://orca.cardiff.ac.uk/id/eprint/152037

Citation Data

Cited 2 times in Scopus. View in Scopus. Powered By Scopus® Data

Actions (repository staff only)

Edit Item