Artículos de revistas
Non-supervised pattern recognition methods: Exploring chemometrical procedures for evaluating analytical data
Registro en:
Quimica Nova. Soc Brasileira Quimica, v. 30, n. 2, n. 481, n. 487, 2007.
0100-4042
WOS:000244752500042
10.1590/S0100-40422007000200042
Autor
Correia, PRM
Ferreira, MMC
Institución
Resumen
An activity for introducing hierarchical cluster analysis (HCA) and principal component analysis (PCA) during the Instrumental Analytical Chemistry course is presented. The posed problem involves the discrimination of mineral water samples according to their geographical origin. Thirty-seven samples of 9 different brands were considered and the results from the determination of Na, K, Mg, Ca, Sr and Ba were taken into account. Non-supervised methods for pattern recognition were explored to construct a dendrogram, score and loading plots. The devised activity can be adopted for introducing Chemometrics devoted to data handling, stressing its importance in the context of modem Analytical Chemistry. 30 2 481 487