As potencialidades da técnica de estatística multivariada análise de componentes principais

Abstract

This work presents a study on the principal components analysis, one of the techniques of multivariate statistics. Initially, a theoretical study on this technique and some existing applications in the literature are simulated in software R. We present three different applications found in articles, highlighting the usefulness of the technique in different contexts. The first application concerns the compression of digital images, another application is presented when the objective is the analysis of the main characteristics of the soil and, a third example, presents the classification of elements of the periodic table. Finally, an application with real data of the instrumentation of concrete dams, in order to emphasize the relevance of the technique regarding the resizing of the data.