DISCRIMINATION OF TUNA (NEOTHYNNUS-ALBACORA) FISHING-SITES USING CHEMICAL-PARAMETERS ELABORATED BY MULTIVARIATE STATISTICAL TECHNIQUES

Data of thirteen parameters (weight, length, weight/length ratio, % dr y weight, % protein, % fat, % ash, Na, K, Ca, Mg, Fe, Cu and Mn) obtai ned from analysis of frozen tuna samples of the same species and size, caught in the Atlantic, Pacific and Indian oceans, were elaborated th rough multivariate statistical techniques: Principal Component Analysi s (PCA), Quadratic Discriminant Analysis (QDA) and Linear Discriminant Analysis (LDA). The results showed the possibility of discriminating between them according to their fishing-site. Use of Feature Selection in order to effect QDA showed that Ca, Mg, Fe and K were the most sta tistically relevant and so they were used for Principal Component Anal ysis which revealed three groups. Classification ability according to Quadratic Discriminant Analysis was about 96%.