Grid-based k Estimation Method for Efficient k-Means Clustering 


Vol. 14,  No. 10, pp. 796-803, Oct.  2025
https://doi.org/10.3745/TKIPS.2025.14.10.796


PDF
  Abstract

As the number of data rapidly increases and its types become more diverse, various types of techniques have been studied to analyze them. Clustering analysis is a representative unsupervised learning technique that partitions a given data into more similar clusters. Since it is difficult to estimate the number of clusters k before performing clustering, previous studies have performed clustering for a certain range of k and then derived the optimal k through score-based evaluation. This has the problem that large-scale calculations are required in a situation where the number of data increases significantly. In this study, to alleviate this problem, we propose a method that reduces the dimension of data, generates grid-based approximate clusters, and estimates the optimal k from the original data set through evaluation using the approximate clusters. The proposed technique was evaluated for performance comparison with the Mini Batch-based method, which is the latest clustering analysis optimization technique, based on a number of synthetic data sets and real data sets, and we demonstrate efficiency in terms of similarity of silhouette scores and reduction of execution time.

  Statistics
  Cite this article

[IEEE Style]

C. Moon, M. H. Kim, J. Min, "Grid-based k Estimation Method for Efficient k-Means Clustering," The Transactions of the Korea Information Processing Society, vol. 14, no. 10, pp. 796-803, 2025. DOI: https://doi.org/10.3745/TKIPS.2025.14.10.796.

[ACM Style]

Cheolhan Moon, Min Hyung Kim, and Jun-Ki Min. 2025. Grid-based k Estimation Method for Efficient k-Means Clustering. The Transactions of the Korea Information Processing Society, 14, 10, (2025), 796-803. DOI: https://doi.org/10.3745/TKIPS.2025.14.10.796.

Forms

Search
(IN TITLE, AUTHOR, ABSTRACT,KEYWORDS)

Advanced Search ▼

POPULAR KEYWORDS
(TOP 10 KEYWORDS)

Recent Publications
(LAST 3 YEARS)

Old Journals

Indexing

All publications of TKIPS is indexed in DOI, EBSCO, Google Scholar, Crossref, and CrossCheck.

TKIPS is also selected as the Journal for Accreditation by NRF (National Research Foundation of Korea).

Related Journals