Malaysian Journal of Mathematical Sciences, March 2026, Vol. 20, No. 1
A Conceptual of Merging of Intuitionistic Fuzzy C-Means with Chebyshev for Genomic Clustering Solutions Addressing Cancer Issues
Zamri, N., Bakar, N. A. A., Aziz, A. Z. A., Madi, E. N., Ramli, R. A., Sukono, Koon, C. S., and Marhadi
Corresponding Email: nadiahzamri@unisza.edu.my
Received date: 18 May 2025
Accepted date: September 2025
Abstract:
Clustering is a fundamental technique for identifying structures within datasets, with Fuzzy C-Means (FCM) being widely used due to its simplicity and ease of implementation. However, FCM suffers from sensitivity to noise, outliers, and initialization issues. This study introduces an enhanced model, IFCM with Chebyshev, which integrates fuzzy Chebyshev distance and intuitionistic fuzzy sets. Data are first normalized using MinMax scaling, and dimensionality reduction is applied to handle high-dimensional datasets. The optimal cluster number is determined using the Elbow method. The proposed algorithm is evaluated against standard FCM, FCM with Chebyshev, and IFCM. A genomic dataset related to prostate cancer is used as a numerical example. Results show that IFCM with Chebyshev achieves the highest clustering accuracy (88.9%), outperforming IFCM (85.7%), FCM with Chebyshev (81.2%), and FCM (78.5%). It also yields superior cluster validity indices, recording the highest Partition Coefficient (0.74) and the lowest Partition Entropy (0.52), indicating clearer cluster separation. Although IFCM with Chebyshev incurs higher computational cost (1.58s), sensitivity analysis demonstrates faster convergence around five clusters, suggesting an optimal structure. Memory consumption remains consistent at 295 MB across cluster settings, highlighting efficiency for large-scale applications. Overall, combining Chebyshev distance and IFCM enhances clustering robustness and accuracy, particularly in noisy or complex data environments, making it a promising approach for advanced data analysis tasks.
Keywords: clustering; FCM; intuitionistic FCM; Chebyshev; prostate cancer; genomic clustering.
