Article Per Year (5 Year)
p-Index From 2021 - 2026
0.23
P-Index
This Author published in this journals
All Journal JITU : Journal Informatic Technology And Communication
Umar Ghoni
Unknown Affiliation
Computer Science & IT Control & Systems Engineering Decision Sciences, Operations Research & Management Library & Information Science

Published : 1 Documents Claim Missing Document
Claim Missing Document
Check
Articles

Found 1 Documents
Search

 1 
Optimizing Clustering Performance: A Novel Integration of Whale Optimization Algorithm and K-NN Validation in Data Mining Analytics Nur Wahyu Hidayat; Umar Ghoni; Mursalim
JITU Vol 9 No 1 (2025)
Publisher : Universitas Boyolali

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.36596/jitu.v9i1.1808

Abstract

The digital era's massive data necessitates effective clustering, a machine learning technique grouping data by similarity. Clustering large, complex datasets faces challenges like volume, dimensionality, and variability, hindering algorithms like K-Means. A key issue in K-Means is its sensitivity to initial centroid selection, impacting results. This research aims to optimize clustering performance by integrating the Whale Optimization Algorithm (WOA) for improved initial centroid determination in K-Means, and K-Nearest Neighbors (K-NN) for validating the resulting cluster quality through classification accuracy. Evaluation on iris, wine, heart, lung, and liver datasets using the Davies-Bouldin Index (DBI) showed that WOA-KMeans consistently yielded lower DBI values compared to standard K-Means, indicating superior clustering. Notably, DBI for the lung dataset drastically decreased from 2.38016 to 0.65395. Furthermore, K-NN classification using the generated cluster labels achieved high accuracy (98-99% across datasets), confirming well-separated and internally homogeneous clusters. This demonstrates WOA's effectiveness in guiding K-Means towards better solutions and K-NN's utility in validating cluster distinctiveness. This novel WOA-K-NN combination offers a more accurate and robust clustering method. The significant performance improvements observed across diverse datasets highlight its potential for enhanced data exploration and pattern discovery in complex data mining tasks.
Co-Authors Mursalim Mursalim Nur Wahyu Hidayat