Ordinal Regression Model using Bootstrap Approach

Abstract

The aim of the research content three part, thus a to know misclassification and model discriminant analysis with bootstrap approach, model regression ordinal with bootstrap approach, and model MARS with bootstrap approach. The data used is data of secondary related to matrix variance covariance is same and unequal that is (The data worker standard of living and banking performance). The result of this research shows that in determining distinguishing variable between groups there are difference of variable at each method. This matter because of at each method has specification either from fulfilled of assumption and also estimation its. So also at accuracy of classification between groups there is difference especially at matrix of variance covariance unequal at worker standard of living case. As a whole can be concluded that the problem accuracy of classification bootstrap approach at each method give small mistake of goodness at matrix variance covariance unequal and equal.

Keywords: classification, bootstrap, discriminant analysis, ordinal regression, MARS.