1. Vectorized Kernel-Based Fuzzy C-Means: a Method to Apply KFCM on Crisp and Non-Crisp Numbers.
- Author
-
Hossein-Abad, Hadi Mahdipour, Shabanian, Mohsen, and Kazerouni, Iman Abaspur
- Subjects
REMOTE-sensing images ,ALGORITHMS ,GENE expression ,SOIL mapping ,GENE clusters ,THEMATIC mapper satellite ,FUZZY numbers ,IMAGE segmentation - Abstract
Kernel methods are a class of algorithms for pattern analysis to robust them to noise, overlaps, outliers and also unequal sized clusters. In this paper, kernel-based fuzzy c-means (KFCM) method is extended to apply KFCM on any crisp and non-crisp input numbers only in a single structure. The proposed vectorized KFCM (VKFM) algorithm maps the input (crisp or non-crisp) features to crisp ones and applies the KFCM (with prototypes in feature space) on them. Finally the resulted crisp prototypes in the mapped space are influenced by an inverse mapping to obtain the prototypes' (centers') parameters in the input features space. The performance of the proposed method has been compared with the conventional FCM and KFCM and other new methods, to show its effectiveness in clustering of gene expression data and segmentation of land-cover using satellite images. Simulation results show good accuracy of proposed method in compare to other methods. [ABSTRACT FROM AUTHOR]
- Published
- 2020
- Full Text
- View/download PDF