武大SAS要点步骤Word格式文档下载.docx
- 文档编号:18051081
- 上传时间:2022-12-13
- 格式:DOCX
- 页数:43
- 大小:255.09KB
武大SAS要点步骤Word格式文档下载.docx
《武大SAS要点步骤Word格式文档下载.docx》由会员分享,可在线阅读,更多相关《武大SAS要点步骤Word格式文档下载.docx(43页珍藏版)》请在冰豆网上搜索。
dataex2_5;
inputxf@@;
y=log10(x);
104
203
4010
8010
16011
32015
64014
12802
procmeansnoprint;
vary;
freqf;
outputout=b
mean=logmean;
datac;
setb;
g=10**logmean;
procprintdata=c;
varg;
例3-2求95%可信区间
dataex3_2;
n=10;
mean=166.95;
std=3.64;
t=tinv(0.975,n-1);
pts=t*std/sqrt(n);
lclm=mean-pts;
uclm=mean+pts;
procprint;
varlclmuclm;
例3-4求两总体相差多大(区间估计)
dataex3_4;
n1=29;
n2=32;
m1=20.10;
m2=16.89;
s1=7.02;
s2=8.46;
ss1=s1**2*(n1-1);
ss2=s2**2*(n2-1);
sc2=(ss1+ss2)/(n1+n2-2);
se=sqrt(sc2*(1/n1+1/n2));
t=tinv(0.975,n1+n2-2);
lclm=(m1-m2)-t*se;
uclm=(m1-m2)+t*se;
vartselclmuclm;
例3-5单样本t检验
dataex3_5;
n=36;
s_m=130.83;
std=25.74;
p_m=140;
df=n-1;
t=abs(s_m-p_m)/(std/sqrt(n));
p=(1-probt(t,df))*2;
vartp;
可以看到,检验统计量t=-2.13753,其对应的P值为0.039618,小于显著性水平的临界值0.05,故拒绝H0,接受H1,即认为样本均数与总体均数在统计学上具有显著性差异。
例3-6配对样本t检验
dataex3_6;
inputx1x2@@;
d=x1-x2;
0.8400.580
0.5910.509
0.6740.500
0.6320.316
0.6870.337
0.9780.517
0.7500.454
0.7300.512
1.2000.997
0.8700.506
procmeanstprt;
vard;
procunivariatedata=ex3_6;
从结果中可以看出检验统计量t值为7.93.对应的p<0.0001,按α=0.05水准,拒绝H0,接受H1,差异有统计学意义,即认为配对资料两样本均数在统计学上具有显著性差异。
例3-7两样本t检验
dataex3_7;
inputx@@;
if_n_<
21thenc=1;
elsec=2;
-0.70-5.602.002.800.703.504.005.807.10-0.50
2.50-1.601.703.000.404.504.602.506.00-1.40
3.706.505.005.200.800.200.603.406.60-1.10
6.003.802.001.602.002.201.203.101.70-2.00
procttest;
varx;
classc;
首先进行方差齐性检验,结果显示统计量F值为1.60,对应额的P值为0.3153,大于显著性水平的临界值0.05,认为两样本具有方差齐性;
然后对两样本均数进行差异性分析,结果显示统计量T值为-0.64,对应的P值为0.5248,大于0.05,接受H0,即认为两样本均数不具有显著性差异。
例4-2完全随机设计资料的方差分析
dataex4_2;
inputxc@@;
3.5312.4222.8630.894
4.5913.3622.2831.064
4.3414.3222.3931.084
2.6612.3422.2831.274
3.5912.6822.4831.634
3.1312.9522.2831.894
3.3012.3623.4831.314
4.0412.5622.4232.514
3.5312.5222.4131.884
3.5612.2722.6631.414
3.8512.9823.2933.194
4.0713.7222.7031.924
1.3712.6522.6630.944
3.9312.2223.6832.114
2.3312.9022.6532.814
2.9811.9822.6631.984
4.0012.6322.3231.744
3.5512.8622.6132.164
2.6412.9323.6433.374
2.5612.1722.5832.974
3.5012.7223.6531.694
3.2511.5623.2131.194
2.9613.1122.2332.174
4.3011.8122.3232.284
3.5211.7722.6831.724
3.9312.8023.0432.474
4.1913.5722.8131.024
2.9612.9723.0232.524
4.1614.0221.9732.104
2.5912.3121.6833.714
procanova;
modelx=c;
meansc/dunnett;
meansc/hovtest;
Pr>
F,方差齐
例4-4随机区组设计资料的方差分析
dataex4_4;
inputxab@@;
0.8211
0.6521
0.5131
0.7312
0.5422
0.2332
0.4313
0.3423
0.2833
0.4114
0.2124
0.3134
0.6815
0.4325
0.2435
classab;
modelx=ab;
meansa/snk;
例7-1四格表卡方检验
dataex7_1;
inputrcf@@;
1199
125
2175
2221
procfreq;
weightf;
tablesr*c
/chisq
expected;
例7-2校正卡方检验
dataex7_2;
1146
126
2118
228
例7-6多个样本率的比较
dataex7_6;
11199
127
21164
2218
31118
3226
/chisq;
例7-7样本构成比的比较
dataex7_7;
1142
1248
1321
2130
2272
2336
例7-8双向无序分类资料的关联性检验
dataex7_8;
11431
12490
13902
21388
22410
23800
31495
32587
33950
41137
42179
4332
例8-1配对样本比较的Wilcoxon符号秩检验
dataex8_1;
6076
142152
195243
8082
242240
220220
190205
2538
198243
3844
236190
95100
procunivariate;
例8-3两个独立样本比较的Wilcoxon秩和检验
dataex8_3;
2.781
3.231
4.201
4.871
5.121
6.211
7.181
8.051
8.561
9.601
3.232
3.502
4.042
4.152
4.282
4.342
4.472
4.642
4.752
4.822
4.952
5.102
procnpar1waywilcoxon;
一、多元回归
1、普通多元回归
dataa;
inputidx1-x4y;
1173106714.7137
21391326.417.8162
31981126.916.7134
41181387.115.7188
5139948.613.6138
617516012.120.3215
713115411.221.5171
81581419.729.6148
91581377.418.2197
101321517.517.2113
11162110615.9145
1214411310.142.881
131621377.220.7185
141691298.516.7157
151291386.310.1197
1616614811.533.4156
17185118617.5156
181551216.120.4154
191751114.127.2144
201361109.42690
211531338.516.9215
221101499.524.7184
23160865.310.8118
24112123816.6127
251471108.518.4137
262041226.121126
271311026.613.4130
2817011278.424.7135
291731238.719188
3013213113.829.2122
procreg;
modely=x1-x4/stb;
结果解释:
根据方差分析结果:
F=4.26,P=0.0092<
0.05,回归方程有统计学意义
根据参数估计结果:
仅X3有意义,t=-2.42,P<
0.05
回归方程是:
=178.47344+0.09028X1-0.00104X2-1.3939X3-1.40112X4
2、多元回归(逐步回归法)
inputidx1-x5y;
1173106714.713762
21391326.417.816243
31981126.916.713481
41181387.115.718839
5139948.613.613851
617516012.120.321565
713115411.221.517140
81581419.729.614842
91581377.418.219756
101321517.517.211337
11162110615.914570
1214411310.142.88141
131621377.220.718556
141691298.516.715758
151291386.310.119747
1616614811.533.415649
17185118617.515669
181551216.120.415457
191751114.127.214474
201361109.4269039
211531338.516.921565
221101499.524.718440
23160865.310.811857
24112123816.612734
251471108.518.413754
262041226.12112672
271311026.613.413051
2817011278.424.713562
291731238.71918885
3013213113.829.212238
modely=x1-x5/stbselection=stepwisesle=0.10sls=0.15;
根据逐步回归的条件:
引入标准=0.10,剔除标准=0.15,回归方程纳入X1和X4,P均小于0.05
由方差分析结果显示:
F=46.48,P<
0.0001,回归方程有统计学意义。
根据参数估计结果显示:
=-11.78059+0.49842X1-0.49666X4
3.控制变量后的相关系数
- 配套讲稿:
如PPT文件的首页显示word图标,表示该PPT已包含配套word讲稿。双击word图标可打开word文档。
- 特殊限制:
部分文档作品中含有的国旗、国徽等图片,仅作为作品整体效果示例展示,禁止商用。设计者仅对作品中独创性部分享有著作权。
- 关 键 词:
- SAS 要点 步骤
![提示](https://static.bdocx.com/images/bang_tan.gif)