Advances in Operations Research

Research Article

Mathematical Programming Approaches to Classification Problems

Table 2

The correct classification rates of the MCA and MSD models (in percentage).
D1D2D3D4
𝑛 1 = 2 1 , 𝑛 1 = 2 2 , 𝑛 1 = 5 0 , 𝑛 1 = 4 4 4 ,
𝑛 2 = 2 5 , 𝑛 2 = 4 0 , 𝑛 2 = 5 0 , 𝑛 2 = 2 3 8 ,
𝑛 = 4 6 , 𝑛 = 6 2 , 𝑛 = 1 0 0 , 𝑛 = 6 8 3 ,
MCAMSDMCAMSDMCAMSDMCAMSD
(N1) ∑ 𝑘 𝑗 = 1 𝑎 𝑗 + 𝑐 = 1 91,384,7883,8764,52968697,294,7
( 4 )( 7 ) ( 1 0 )( 2 2 ) ( 4 ) ( 1 4 )( 1 9 )( 3 6 )
(N2) ∑ 𝑘 𝑗 = 1 𝑎 𝑗 = 1 91,389,183,8766,13969197,296,6
( 4 )( 5 ) ( 1 0 )( 2 1 ) ( 4 ) ( 9 )( 1 9 )( 2 3 )
(N3) 𝑐 ± 1 91,389,183,8772,6969197,296,6
( 4 )( 5 ) ( 1 0 )( 1 7 ) ( 4 ) ( 9 )( 1 9 )( 2 3 )
(N4) invariance under91,386,9583,8775,8969197,296,6
origin shift ( 4 )( 6 ) ( 1 0 )( 1 5 ) ( 4 )( 9 )(1 9 )( 2 3 )
The values between parentheses are the numbers of misclassified observations.