ABSTRACT
The aim of this study was to verify if the principal components method of analysis could be employed to characterize commercial powdermainly starch used in pudding preparation, identifying and reducing the most important factors. The multivariate analysis was applied to data regarding composition and functional properties which were taken from twelve samples totaling fourteen variables. The data were analyzed using a personal computer system ARTHUR. Seven variables (starch, fracturability, firmness, adhesiviness, cohesiveness, elasticity, total solids) of greater correlation and higher absolute values that were characteristic of the two principal components selected after the first examination. A second analysis which used only selected variables maintained the initial behaviour confirming the possibility of reducing variables without important loss of information. The application of the multivariate analysis using the principal components method for the study of commercial puddings proved to be efficient in the classification based on sample similarity, besides selecting and reducing variables to more outstanding factors