42=1=(5.68x1.90+3.79x1.64+.+3.84x1.20- 5.68+3.79+.+3.841.90+1.64+.+1.20) 27 =513.1867-156.94x76.7/27 =67.3608
各变量的离差矩阵 X1 X2 X3 X4 X166.0103 67.3608 -53.9523 31.3687 67.6962 X2 67.3608 172.3648 -9.4929 26.7286 89.8025 1 X3-53.9523 -9.4929 350.3106-57.3863-142.4347 X431.3687 26.7286 -57.3863 86.4407 84.5570 761.6962 89.8025 -142.4347 84.5570 222.5519
各变量的离差矩阵 6.0103%+67.360862-53.9523%3+31.36876=67.6962 67.3608+1723648%,-9.492%2+26.7286,=89.8025 -53,9523%-9.4929%2+350.3103-57.38636,=-1424347 31.36876+26.72865,-57.38635+86.440764=84.5570
建立多元回归方程 求解后得 b=0.1424 b2=0.3515 b,=-0.2706 b,=0.6382 算得均值分别为元1=5.8126，x2=2.8407,3=6.1467,x4=9.1185, 了-11.9259，按照公式b0=7-(dz+b22+b3x3+b44)可求得常数项 b=11.9259-(0.1424×5.8126+0.3515×2.8407-0.2706×6.1467+0.6382×9.1185) =5.9433 故所求多元回归方程为 立=5.9433+01424X,+0.3515X2-0.2706X3+0.6382X
三、多元线性回归方程的 假设检验及其评价