上机课第8讲 判别分析过程.ppt

Pooled Within-Class Standardized Canonical Coefficients Variable Can1 Can2 x1 0.4531720478 -.1749855680 x2 0.5964259477 -.8107819809 x3 0.6615904861 0.6001270597 x4 0.2992055421 0.6078416327 Raw Canonical Coefficients Variable Can1 Can2 x1 0.0100479384 -.0038798602 x2 0.0401775618 -.0546174144 x3 0.1763979985 0.1600101791 x4 0.0305481466 0.0620591289 Class Means on Canonical Variables type2 Can1 Can2 n -1.262949592 -0.472278377 w -0.935932739 0.521545394 y 2.198882331 -0.049267017 Obs type1 type2 x1 x2 x3 x4 Can1 Can2 1 y y 228 134 20 11 2.75459 1.12712 2 y y 245 134 10 40 2.04732 1.26078 3 y y 200 167 12 27 2.87670 -0.85375 4 y y 170 150 7 8 0.42984 -1.78803 5 y y 100 167 20 14 2.88596 0.00755 6 n w 225 125 7 14 0.16132 -0.26364 7 n w 130 100 6 12 -2.03517 1.18626 8 n w 150 117 7 6 -1.15808 -0.03218 9 n w 120 133 10 26 0.32348 0.93155 10 n w 160 100 5 10 -1.97122 0.78573 11 n n 185 115 5 19 -0.84243 0.42801 12 n n 170 125 6 4 -0.87319 -0.83084 13 n n 165 142 5 3 -0.44736 -1.96201 14 n n 135 108 2 12 -2.36910 0.08988 15 n n 100 117 7 2 -1.78267 -0.08642 先将分析变量中心化，然后得到典型变量的得分，即典型变量的均值为0. The DISCRIM Procedure Linear Discriminant Function Constant = -.5 Xj COV-1Xj Coefficient Vector = COV