วิธีโปรแกรมเชิงเส้นสำหรับปัญหาการแบ่งกลุ่มในการวิเคราะห์การจำแนก

Abstract

The classification problem in discriminant analysis involves the construction a rule tbr the assignment of an observation into one of several groups. Various discriminant analysis techniques have been proposed in the llelds of statistics. management science and artificial intelligence. This paper introduces two most witlely used linear programming (LP) tbrmulations to the classification problem. The results of experimental comparisons of LP and the statistical approaches are reported'