多元分析SAS上機實驗修改ppt課件

1、2022-4-2212022-4-222Y0102030405060708090100110120130140150X0102030405060702022-4-223Model: MODEL1Dependent Variable: YAnalysis of Variance Sum of MeanSource DF Squares Square F Value ProbFModel 1 14141.74052 14141.74052 140.992 0.0001Error 6 601.80823 100.30137C Total 7 14743.54875 Root MSE 10.01506

2、 R-square 0.9592 Dep Mean 43.51250 Adj R-sq 0.9524C.V. 23.016511)1(222 pnRpRRjdA為校正后的決定系數(shù)為校正后的決定系數(shù). 2022-4-224 Parameter Estimates Parameter Standard T for H0:Variable DF Estimate Error Parameter=0 Prob |T|INTERCEP 1 5.922366 4.74969580 1.247 0.2589 X 1 2.076803 0.17490302 11.874 0.0001Model Crosspr

3、oducts XX XY YYXX INTERCEP X YINTERCEP 8 144.8 348.1X 144.8 5899.66 13109.99Y 348.1 13109.99 29890.25XX Inverse, Parameter Estimates, and SSE INTERCEP X YINTERCEP 0.2249182623 -0.005520346 5.9223664747 X -0.005520346 0.0003049915 2.0768029572 Y 5.9223664747 2.0768029572 601.808229322022-4-225Model:

4、MODEL1Dependent Variable: YAnalysis of Variance Sum of MeanSource DF Squares Square F Value ProbFModel 1 14141.74052 14141.74052 140.992 0.0001Error 6 601.80823 100.30137C Total 7 14743.54875 Root MSE 10.01506 R-square 0.9592 Dep Mean 43.51250 Adj R-sq 0.9524C.V. 23.016511)1(222 pnRpRRjdA為校正后的決定系數(shù)為校

5、正后的決定系數(shù). 2022-4-226 Parameter Estimates Parameter Standard T for H0:Variable DF Estimate Error Parameter=0 Prob |T|INTERCEP 1 5.922366 4.74969580 1.247 0.2589 X 1 2.076803 0.17490302 11.874 0.0001Model Crossproducts XX XY YYXX INTERCEP X YINTERCEP 8 144.8 348.1X 144.8 5899.66 13109.99Y 348.1 13109.9

6、9 29890.25XX Inverse, Parameter Estimates, and SSE INTERCEP X YINTERCEP 0.2249182623 -0.005520346 5.9223664747 X -0.005520346 0.0003049915 2.0768029572 Y 5.9223664747 2.0768029572 601.808229322022-4-227 Dep Var Predict Std Err Lower95% Upper95%Obs Y Value Predict Predict Predict Residual 1 46.3000 3

7、5.6206 3.603 9.5773 61.6640 10.6794 2 30.7000 34.9976 3.613 8.9459 61.0493 -4.2976 3 144.6 149.8 9.630 115.8 183.8 -5.2448 4 69.2000 53.0658 3.631 26.9988 79.1327 16.1342 5 16.0000 21.0830 4.013 -5.3172 47.4833 -5.0830 6 12.3000 22.5368 3.957 -3.8127 48.8863 -10.2368 7 2.7000 8.6222 4.601 -18.3464 3

8、5.5908 -5.9222 8 26.3000 22.3291 3.965 -4.0274 48.6856 3.9709 9 . 47.4584 3.556 21.4532 73.4636 . Sum of Residuals 0 Sum of Squared Residuals 601.8082 Predicted Resid SS (Press) 5610.21602022-4-228OBS X Y PRESIDI 1 14.3 46.3 12.2667 2 14.0 30.7 -4.9405 3 69.3 144.6 -69.4832 4 22.7 69.2 18.5761 5 7.3

9、 16.0 -6.0554 6 8.0 12.3 -12.1305 7 1.3 2.7 -7.5067 8 7.9 26.3 4.7089)(iyyPRESIDiii 是第個觀測值的預(yù)測殘差是第個觀測值的預(yù)測殘差,而而 )(iyi是刪去第是刪去第i個觀測值后個觀測值后,根據(jù)剩下的根據(jù)剩下的n1組觀測值建立線組觀測值建立線性回歸方程并將性回歸方程并將xi代入此回歸方程所得到的第代入此回歸方程所得到的第i個觀測個觀測值的預(yù)測值值的預(yù)測值. 2022-4-229多元線性回歸多元線性回歸上機實驗上機實驗2. 例例1.2 P25data lreg07;input x1 x2 y ;cards;1.37

10、9.08 4.93 11.34 1.89 1.86 9.673.06 2.33 0.76 10.2 5.78 17.67 0.050.06 15.91 0.73 0.43 15.74 1.03 0.875.41 6.25 3.86 5.4 6.3 .;symbol i=rl v=star;/*（symbol 為繪圖設(shè)置語句為繪圖設(shè)置語句*/* i=rl 要求畫回歸直線要求畫回歸直線*/*v=star 要求用星號表示點）要求用星號表示點）*/2022-4-2210proc gplot;plot y*x1;/*（要求畫回歸直線及點（要求畫回歸直線及點x1,y）*/proc reg;model y=

11、x1/xpx i;run;/*（要求建立一元線性回歸方程并作檢驗）（要求建立一元線性回歸方程并作檢驗）*/ proc gplot;plot y*x2;proc reg;model y=x2;run;proc reg; model y=x1 x2/alpha=0.01 cli;run;/*（要求建立二元線性回歸方程（要求建立二元線性回歸方程 并作檢驗及回報與預(yù)報）并作檢驗及回報與預(yù)報）*/2022-4-2211Y0123456X1024681012141618Y0123456X2012345678910 112022-4-2212Y=5.621170 - 0.319113*X1;F=494.06

12、4;ProbF為為0.0001;R-square=0.9880;Y=0.363778 + 0.532975*X2;F=279.602; ProbF為為0.0001;R-square=0.9790;Y=3.610506 - 0.198281*X1 + 0.206755*X2;F=392.516; ProbF為為0.0001;R-square=0.9937;2022-4-2213Model Crossproducts XX XY YYXX INTERCEP X1 X2 YINTERCEP 8 77.87 32.29 20.12X1 77.87 1066.9317 33.7369 97.2483X2

13、 32.29 133.7369 40.0809 .7036Y 20.12 97.2483 .7036 82.4468 XX Inverse, Parameter Estimates, and SSE INTERCEP X1 X2 YINTERCEP 22.8232 -1.3771 2.3025 3.6105X1 -1.3771 0.0841 0.1384 -0.1983X2 -2.3025 0.1384 0.2368 0.2068Y 3.6105 -0.1983 0.2068 0.20192022-4-2214 Analysis of Variance Sum of MeanSource DF

14、 Squares Square F Value Prob FModel 2 31.64346 15.82173 392.516 0.0001Error 5 0.20194 0.04031C Total 7 31.84500Root MSE 0.20197 R-square 0.9937Dep Mean 2.51500 Adj R-sq 0.9911C.V. 7.98289 1)1(*:222 pnRpRRqsRjdA2022-4-2215 Dep Var Predict Std Err Lower95% Upper95%Obs Y Value Predict Predict Predict R

15、esidual 1 4.9300 5.2162 0.120 4.6150 5.8174 -0.2862 2 1.8600 1.7528 0. 1.1252 2.3803 0.1072 3 2.3300 2.3258 0.122 1.7221 2.9295 0.00420 4 5.7800 5.5687 0.154 4.9188 6.2186 0.2113 5 0.0600 0.1172 0.132 -0.4998 0.7343 -0.0572 6 0.4300 0.6068 0.104 0.0251 1.1884 -0.1768 7 0.8700 0.7025 0.110 0.1137 1.2

16、914 0.1675 8 3.8600 3.8300 0.092 3.2626 4.3974 0.0300 9 . 3.8423 0.090 3.2763 4.4084 . 觀測值觀測值 預(yù)測值預(yù)測值 標準誤標準誤 95%置信下限與上限置信下限與上限 殘差剩余）殘差剩余）2022-4-2216例例1.4 逐步回歸逐步回歸data ex; input x1-x4 y ;cards;9.2 2.73 1.47 1.14 1.16 9.1 3.73 1.82 0.83 1.15 8.6 4.88 1.83 2.13 1.84 10.23 3.97 1.59 1.35 1.36 5.6 3.73 1.

17、84 1.82 0.86 5.37 4.24 1.87 1.35 0.9 6.13 3.15 1.99 1.65 0.12 8.2 4.65 1.62 4.57 0.9 8.8 4.38 1.54 2.07 1.93 7.6 3.86 1.6 2.42 1.1 9.7 4.38 1.69 1.52 1.4 8.37 5.1 1.81 2.28 1.76 12.17 4.9 1.73 1.58 1.64 10.27 3.73 1.61 1.2 1.47 8.9 4.47 1.88 0.8 0.92 8.23 5.28 1.7 3.07 1.5;proc reg;model y=x1-x4/sel

18、ection=rsquare b;model y=x1-x4/selection=rsquare adjrsq cp AIC MSE SSE collinoint;run;proc reg;model y=x1-x4/selection=stepwise sle=0.3 sls=0.2; run; / * sls is the level spcified by user*/SLS=0.2定義剔除時的概率水平為定義剔除時的概率水平為0.2，SLE=0.3定義選定義選入時的概率水平為入時的概率水平為0.32022-4-2217例例1.4 共線性診斷共線性診斷data ex; input x1-x

19、4 y ; x5=int(x2+x3+x4)*90)/90;cards;9.2 2.73 1.47 1.14 1.16 9.1 3.73 1.82 0.83 1.15 8.6 4.88 1.83 2.13 1.84 10.23 3.97 1.59 1.35 1.36 5.6 3.73 1.84 1.82 0.86 5.37 4.24 1.87 1.35 0.9 6.13 3.15 1.99 1.65 0.12 8.2 4.65 1.62 4.57 0.9 8.8 4.38 1.54 2.07 1.93 7.6 3.86 1.6 2.42 1.1 9.7 4.38 1.69 1.52 1.4 8

20、.37 5.1 1.81 2.28 1.76 12.17 4.9 1.73 1.58 1.64 10.27 3.73 1.61 1.2 1.47 8.9 4.47 1.88 0.8 0.92 8.23 5.28 1.7 3.07 1.5;proc reg data=ex;model y=x1-x5/vif collinoint;run;2022-4-2218Analysis of Variance Sum of Mean Source DF Squares Square F Value Pr F Model 5 2.55497 0.51099 8.03 0.0028 Error 10 0.63

21、672 0.06367 Corrected Total 15 3.19169 Root MSE 0.25233 R-Square 0.8005 Dependent Mean 1.25063 Adj R-Sq 0.7008 Coeff Var 20.17663Parameter Estimates Parameter Standard Variance Variable DF Estimate Error t Value Pr |t| Inflation Intercept 1 2.82510 1.33994 2.11 0.0612 0 x1 1 0.02693 0.06221 0.43 0.6

22、743 2.85303 x2 1 -27.00023 25.70336 -1.05 0.3182 76421 x3 1 -29.73236 25.68616 -1.16 0.2740 3263.98863 x4 1 -27.81537 25.68282 -1.08 0.3042 137920 x5 1 27.58457 25.69586 1.07 0.3083 3083662022-4-2219Collinearity Diagnostics (intercept adjusted) Condition Number Eigenvalue Index 1 2.45588 1.00000 2 1.48451 1.28621 3 0.89498 1.65652 4 0.16462 3.86241 5 0.00000190 1136.53374 Collinearity Diagnostics (int