On the classification gap in mathematical programming-based approaches to the discriminant problem

摘要: This article proposes a mathematical-programming-based approach to solve the classification problem in discriminant analysis which explicitly considers gap. The procedure consists of two distinct phases and initially treats gap as fuzzy set rule is not yet established. nature examined variety methods are discussed can be applied identify most appropriate over set. proposed methodology has several potential advantages. First, it offers more refined problem, facilitating careful region where decision may obvious. Secondly, two-phase enables larger data sets when using computer-intensive procedures such mixed-integer programming. Finally, because restricted choice separating hyperplanes phase 2, appears robust than other techniques with respect outlier-contaminated conditions. robustness issue computational advantage our illustrated limited simulation experiment.