Akademik Veri Yönetim Sistemi
Applied Soft Computing, cilt.195, 2026 (SCI-Expanded, Scopus)
This study examines live specimens of Verbascum speciosum, V. vulcanicum, and V. lasianthum during the vegetation periods, measuring 75 individuals per species for stamen length (mm), number of flowers (count), flower length (mm), calyx (mm), and capsule (mm). The collected specimens are identified and preserved as herbarium materials in the Aksaray University Herbarium (AKSU). To analyze and classify the morphological data under uncertainty arising from natural intraspecific variation and morphological overlap among closely related species, a Spherical Fuzzy C-Means (SFCM) clustering algorithm is developed. Unlike classical fuzzy and intuitionistic fuzzy approaches, spherical fuzzy sets provide a more comprehensive representation by independently defining degrees of membership, non-membership, and neutrality in the central unit sphere. By utilizing this extended fuzzy framework, the proposed SFCM algorithm improves both the reliability and interpretability of clustering outcomes, offering a robust and accurate tool for morphological differentiation and taxonomic classification of plant species. The results are evaluated using clustering validity indices, including partition coefficient, modified partitional coefficient, partition entropy, clustering accuracy, Rand index, adjusted Rand index, and clustering purity. The SFCM algorithm successfully differentiated three morphologically similar Verbascum taxa based primarily on subtle variations in capsule and calyx dimensions according to partition coefficient, modified partitional coefficient, and partition entropy. The diagnostic importance of these reproductive structures is consistent with features repeatedly emphasized in classical taxonomic studies, thereby aligning statistical outcomes with natural biological distinctiveness.