SAS考试
姓名：宋亚琼专业：营养与食品卫生学学号：12720306 第一题：程序
options nodate nonumber;
data zhang1;
input a b c d x@@;
cards;
1 1 1 1 78.9 1 1 1 1 78.1
1 2 2 2 77.0 1 2 2 2 77.0
1 3 3 3 77.5 1 3 3 3 78.5
2 1 2
3 80.1 2 1 2 3 80.9
2 2
3 1 77.6 2 2 3 1 78.4
2 3 1 2 78.0 2 3 1 2 79.0
3 1 3 2 76.7 3 1 3 2 76.3
3 2 1 3 81.3 3 2 1 3 82.7
3 3 2 1 79.5 3 3 2 1 78.5
;
proc anova;
class a b c d;
model x=a b c d;
means a b c d;
means a b c d/duncan;
means a b c d/duncan alpha=0.01;
run;
SAS 系统
The ANOVA Procedure
Class Level Information
Class Levels Values
a 3 1 2 3
b 3 1 2 3
c 3 1 2 3
d 3 1 2 3
Number of observations 18
SAS 系统
The ANOVA Procedure
Dependent Variable: x
Source DF Sum of Squares Mean Square F Value Pr > F
Model 8 46.00000000 5.75000000 14.70 0.0003 Error 9 3.52000000 0.39111111
Corrected Total 17 49.52000000
R-Square Coeff Var Root MSE x Mean
0.928918 0.794986 0.625389 78.66667
Source DF Anova SS Mean Square F Value Pr > F
a 2 6.33333333 3.16666667 8.10 0.0097
b 2 1.00000000 0.50000000 1.28 0.3246
c 2 14.33333333 7.16666667 18.32 0.0007
d 2 24.33333333 12.16666667 31.11 <.0001
SAS 系统
The ANOVA Procedure
Level of --------------x--------------
a N Mean Std Dev
1 6 77.8333333 0.79414524
2 6 79.0000000 1.27593103
3 6 79.1666667 2.52560224
Level of --------------x--------------
b N Mean Std Dev
1 6 78.5000000 1.82866071
2 6 79.0000000 2.42074369
3 6 78.5000000 0.70710678
Level of --------------x--------------
c N Mean St
d Dev
1 6 79.6666667 1.90438091
2 6 78.833333
3 1.62193300
3 6 77.5000000 0.88317609
Level of --------------x--------------
d N Mean Std Dev
1 6 78.5000000 0.65421709
2 6 77.333333
3 0.99129545
3 6 80.1666667 1.90438091
Duncan's Multiple Range Test for x
NOTE: This test controls the Type I comparisonwise error rate, not the experimentwise error rate.
Alpha 0.05
Error Degrees of Freedom 9
Error Mean Square 0.391111
Number of Means 2 3
Critical Range .8168 .8525
Means with the same letter are not significantly different.
Duncan Grouping Mean N a
A 79.1667 6 3
A
A 79.0000 6 2
B 77.8333 6 1
Duncan Grouping Mean N b
A 79.0000 6 2
A
A 78.5000 6 1
A
A 78.5000 6 3
Duncan Grouping Mean N c
A 79.6667 6 1
B 78.8333 6 2
C 77.5000 6 3
Duncan Grouping Mean N d
A 80.1667 6 3
B 78.5000 6 1
C 77.3333 6 2
Alpha 0.01
Error Degrees of Freedom 9
Error Mean Square 0.391111
Number of Means 2 3
Critical Range 1.173 1.222
Means with the same letter are not significantly different.
Duncan Grouping Mean N a
A 79.1667 6 3
A
B A 79.0000 6 2
B
B 77.8333 6 1
Duncan Grouping Mean N b
A 79.0000 6 2
A
A 78.5000 6 1
A
A 78.5000 6 3
Duncan Grouping Mean N c
A 79.6667 6 1
A
A 78.8333 6 2
B 77.5000 6 3
Duncan Grouping Mean N d
A 80.1667 6 3
B 78.5000 6 1
B
B 77.3333 6 2
结果分析：
1、方差分析结果表明：
茶园施肥3要素配合比例（A）、制茶工艺流程（C）、施肥用量（D）均达到的极显著水平(p<0.01)，说明它们对提高炒青绿茶品质有显著效果，而鲜叶处理（B）对提高炒青绿茶品质影响不明显（P>0.05）。
2、用Duncan多重比较：鲜叶处理（B）三水平在P=0.05，P=0.01水平均无差异。
在P=0.05水平：茶园施肥3要素配合比例（A）水平1与2差异不显著，3与1,2差异显著。制茶工艺流程（C）和施肥用量（D）各水平两两间差异显著。因此最优组合：A3C1D3，B因素依据实际情况节约成本选
择配比。
在P=0.01水平：茶园施肥3要素配合比例（A）水平1、2差异不显著，水平2、3差异不显著，水平1、3差异显著。制茶工艺流程（C）水平1、2与3差异显著。施肥用量（D）水平1与2、3差异显著。因此最优组合：A3C1D3，B因素三水平任选其一。
即：提高炒青绿茶品质最优组合是：茶园施肥3要素配合比例水平3，制茶工艺流程水平1，施肥用量水平3。鲜叶处理（B）对提高炒青绿茶品质无影响。
第二题:
程序：
options nodate nonumber;
data zhang2;
input x1 x2 x3 y@@;
x4=x1*x2;x5=x1*x3;
x6=x2*x3;x7=x1*x1;
x8=x2*x2;x9=x3*x3;
cards;
380 140 1.7 4800
380 140 1.3 5050
380 110 1.7 3900
380 110 1.3 4750
320 140 1.7 3900
320 140 1.3 3700
320 110 1.7 3700
320 110 1.3 3850
400 125 1.5 5000
300 125 1.5 3600
350 150 1.5 4400
350 100 1.5 3950
350 125 1.8 4700
350 125 1.2 4750
350 125 1.5 4850
350 125 1.5 4550
350 125 1.5 4400
350 125 1.5 4300
350 125 1.5 4750
350 125 1.5 4350
;
proc reg;model y=x1-x9/selection=backward sls=0.05;run;
输出结果：
SAS 系统
Model: MODEL1
Dependent Variable: y
Backward Elimination: Step 0
All Variables Entered: R-Square = 0.9028 and C(p) = 10.0000
Analysis of Variance
Source DF Sum of Squares MeanSquare F Value Pr > F
Model 9 3822975 424775 10.33 0.0006
Error 10 411400 41140
Corrected Total 19 4234375
arameter Standard
Variable Estimate Error Type II SS F Value Pr > F
Intercept -19699 14014 81290 1.98 0.1901
x1 107.65217 49.84095 191927 4.67 0.0561
x2 28.18382 89.48203 4081.23636 0.10 0.7593
x3 1231.85680 7001.68111 1273.44333 0.03 0.8639
x4 0.31944 0.15936 165313 4.02 0.0728
x5 -23.95833 11.95186 165312 4.02 0.0728
x6 39.58333 23.90373 112813 2.74 0.1287
x7 -0.13953 0.05998 222661 5.41 0.0423
x8 -0.75811 0.23990 410836 9.99 0.0102
x9 585.21440 1581.64999 5632.13527 0.14 0.7191
Bounds on condition number: 736.66, 34697
------------------------------------------------------------------------------------------------
Backward Elimination: Step 1
Variable x3 Removed: R-Square = 0.9025 and C(p) = 8.0310
Analysis of Variance Source DF Sum of Squares MeanSquare F Value Pr > F
Model 8 3821702 477713 12.73 0.0001
Error 11 412673 37516
Corrected Total 19 4234375
Parameter Standard
Variable Estimate Error Type II SS F Value Pr > F
Intercept -18224 10726 108313 2.89 0.1174
x1 105.55552 46.21447 195713 5.22 0.0432
x2 25.18861 83.88897 3382.29882 0.09 0.7696
x4 0.31944 0.15218 165313 4.41 0.0597
x5 -22.70203 9.15238 230821 6.15 0.0305
x6 41.37805 20.64365 150723 4.02 0.0703
x7 -0.13923 0.05725 221878 5.91 0.0333
x8 -0.75690 0.22900 409861 10.93 0.0070
x9 773.79525 1110.65559 18210 0.49 0.5004
Backward Elimination: Step 2
Variable x2 Removed: R-Square = 0.9017 and C(p) = 6.1132
Analysis of Variance
Source DF Sum of Squares MeanSquare F Value Pr > F
Model 7 3818320 545474 15.73 <.0001
Error 12 416055 34671
Corrected Total 19 4234375
Parameter Standard
Variable Estimate Error Type II SS F Value Pr > F
Intercept -16013 7496.69573 158195 4.56 0.054
x1 101.89259 42.85214 196024 5.65 0.0349
x4 0.34846 0.11302 329546 9.50 0.0095
x5 -23.09881 8.70638 244047 7.04 0.0211
x6 43.60870 18.51601 192319 5.55 0.0364
x7 -0.13832 0.05496 219616 6.33 0.0271
x8 -0.71032 0.16193 667116 19.24 0.0009
x9 727.36029 1057.31947 16408 0.47 0.5046
Backward Elimination: Step 3
Variable x9 Removed: R-Square = 0.8979 and C(p) = 4.5120
Analysis of Variance
Source DF Sum of Squares Mean Square F Value Pr > F Model 6 3801912 633652 19.05 <.0001 Error 13 432463 33266
Corrected Total 19 4234375
Parameter Standard
Variable Estimate Error Type II SS F Value Pr > F
Intercept -15230 7257.98822 146473 4.40 0.0560
x1 97.56637 41.52052 183688 5.52 0.0352
x4 0.33947 0.10997 317004 9.53 0.0087
x5 -18.77911 5.90745 336168 10.11 0.0073
x6 48.91332 16.48941 292718 8.80 0.0109
x7 -0.13979 0.05379 224655 6.75 0.022
x8 -0.72952 0.15625 725188 21.80 0.0004
All variables left in the model are significant at the 0.0500 level.
Summary of Backward Elimination
Variable Number Partial Model
tep Removed Vars In R-Square R-Square C(p) F Value Pr > F
1 x3 8 0.0003 0.9025 8.0310 0.03 0.863
2 x2 7 0.0008 0.9017 6.1132 0.09 0.7696
3 x9 6 0.0039 0.8979 4.5120 0.47 0.5046
第二题第三小题程序：
data tan;
do x1=300 to 400 by 2;
do x2=100 to 150 by 1;
do x3=1.2 to 1.8 by 0.05;
y=-15230+97.56637*x1+0.33947*x1*x2-18.77911*x1*x3+48.91332*x2*x3-0.1397 9*x1*x1-0.72952*x2*x2;
output;end;end;end;
proc rank descending;var y;ranks ry;
proc print;
proc g3d;
plot x1*x2=y;plot x1*x3=y;plot x2*x3=y;
run;
结果分析：
（1）本试验采用二次旋转设计，逐步淘汰法，剔除标准为p=0.05,最终去除了x9,x2,x3三个差异不显著因素，最终留下6个变量，得到二次式方程，即：
y=-15230+97.56637*x1+0.33947*x1*x2-18.77911*x1*x3+4
8.91332*x2*x3-0.13979*x1*x1-0.72952*x2*x2;
此方程的拟合效果最好（F= 19.05 ，R2=0.8979，p<0.01）。
（2）依据最优回归方程，进行计算机模拟试验，列出10个最高产量水平组合：
（3）