An application of linear programming discriminated analysis for classification

Abstract EN

The goal of this study is to compare linear discrimination analysis and discriminated analysis with linear programming (MMD) (min.

sum of deviation) in order to find the best model for classifying observations into their correct groups with the lowest possible classification error and highest classification accuracy.

according to the findings of the study, discriminated analysis using linear programming differs from linear discriminated analysis in data classification because it produces the lowest error rate and the highest classification accuracy rate, and it does not require the linear discriminated analysis assumptions.