PRINCIPAL COMPONENT ANALYSIS FOR SELECTION OF VARIABLES IN HOMOGENEITY AND STABILITY TESTS APPLIED TO PROFICIENCY TESTING SCHEMES.

Proficiency Tests (PT) based on interlaboratory comparisons are activities aimed at assessing the technical competence of laboratories in carrying out specific measurements. The analyses of homogeneity and stability of prepared samples are an important step in ensuring the reliability of the comparison rounds, since improper selection of the parameter to carry out this evaluation can influence the promoted comparison. This paper proposes a method for selecting the most relevant variables aimed at improving homogeneity and stability tests in PT. For that matter, the approach relies on a variable importance index derived from Principal Components Analysis (PCA) parameters. The proposed method was applied to three different PT schemes (beverage, water and coal) in Brazil. Results indicate that the use of PCA was adequate to help the variable selection of homogeneity and stability tests in PT schemes. The selected subset of variables was corroborated by experts in the PT schemes analyzed. [ABSTRACT FROM AUTHOR]