Statistical model using discriminant analysis to classify ocular refractive problems in the city of Barranquilla

Abstract

The present work intends to build a multivariate statistical model to classify patients with ocular refractive problems (hyperopia, myopia, astigmatism, hyperopic astigmatism and myopic astigmatism) in terms of a set of variables that specify a type of eye refractive defect. This study was developed in the North of Barranquilla, analyzing the medical histories provided by a clinic in Optometry. In total 499 patients were selected at random to analyze their diagnostic. Their ages were in the range of 8 to 82, with different professions. The database has 15 quantitative variables and 5 qualitative variables. The Multivariate analyzing techniques than were applied to the study, were the principal components factorial analysis with the aim of creating new variables than summarize all information which it may have in the original variables and the discriminant analysis with the purpose of classifying patients in some of the groups defined by the type of defects refractive manifested in his diagnostic.