Biological taxonomic problem solving using fuzzy decision-making analytical tools

Biological taxonomy is at the heart of species identifications. Such identifications are instrumental in biodiversity studies, ecological assessment, and phylogenetic analysis, among other studies. Fuzzy measures and classification integration was used to analyze shape groups of the diatom Asterionella using fuzzy Fourier shape coefficients and fuzzy morphometric measures. Based on this analysis, six shape groups were determined with specimen membership assignments at or exceeding the crossover point (0.5). Fuzzy average overlap values were approximately at or just over the crossover point, indicating similarity in developmental stages. In further analysis, spatial and temporal data from specimen samples were used in conjunction with fuzzy membership assignment values. Spatial and temporal variables were ranked and fuzzified based on the mode. The modes were then weighted by degree of importance as determined by an expert in diatom research. The weighted fuzzy modes for each specimen in each shape group were aggregated as a weighted sum. The normalized relative cardinality for each specimen defined the degree of suitability that a specimen belonged to a shape group, and the expert evaluated the result. While morphological data specifies inheritance (shape) and development (morphometry), spatial and temporal data were proxies for reproductive isolation. These biological principles constrained and defined the direction of analysis and defined each shape group as a species to the degree specified by each specimen. This fuzzy decision-making process provided a simple way to aggregate scant available data and a linguistic solution in a taxonomic study understandable to a biologist. © 2006 Elsevier B.V. All rights reserved.