(己解決) SAS v.s. law of large number

是 否 +2 论坛币

k人 参与回答

经管之家送您一份

应届毕业生专属福利!

求职就业群

赵安豆老师微信：zhaoandou666

经管之家联合CDA

送您一个全额奖学金名额~ !

立即领取

感谢您参与论坛问题回答

经管之家送您两个论坛币！

+2 论坛币

我之前是用R寫過大數法則以及***極限定理的小程式,但,在SAS上遇到困難,
我目前寫的程式如下:
DM LOG 'CLEAR';
DM OUTPUT 'CLEAR';

DATA one;
ARRAY k {100} k1-k100;
   DO n=1 TO 20;
   i=1;
       DO i=1 to 100;
       k(i)=RANNOR(i);
       END;
   OUTPUT;
   END;
RUN;
PROC PRINT NOOBS DATA=one;
RUN;

上半部似乎是沒有問題,因為可以print出來;但,下半部分,我想要把之前建立的one 中的k1 到 k100 combine 成一個k 矩陣,row=20,column=100 可以利用迴圈來處理嗎?

PROC IML;
/*read data file into iml*/
use one;
read all;
k=k1|| k2 || k3 || k4;  (希望可以一直combine到k100,但我只會手動,不會利用迴圈combine)
xmean=k[:,];  /*此部分是想要計算平均值*/
之後我還想要畫一個直條圖,但,在PROC IML下要如何畫呢?

RUN;


希望高手可以幫幫忙教我.


以上的部分為law of large  number....至於central limit theorem我還沒開始寫程式碼,
我想先把這部分弄懂後,也許就會寫出來了...
麻煩大家了...
p.s.
附檔類似的R程式碼,希望能幫助你們了解我要寫的SAS程式碼是什麼

T<-50
# T is sample size
N<-5000
# N is Replication number
Xmean<-matrix(nr=N,nc=1)
for (i in 1:N)
{

X<-rnorm(T,mean=0,sd=1)
Xmean[i,1]<-mean(X)
}
hist(Xmean)

扫码加我 拉你入群

请注明：姓名-公司-职位

以便审核进群资格，未注明则拒绝

分享0 收藏0 回帖

关键词：Number Large arge Ber NUM number file

1# sherryshin

1. 
   
2. DATA one;
   
3.    ARRAY n{20} n1-n20; * a row represents a sample of size 20;
   
4.    do sample =1 to 100; *generate 100 samples;
   
5.        do n= 1 to 20;
   
6.           n[n] = rannor(1);
   
7.       end;
   
8.       mean = mean(of n1-n20); *100 sample means computed;
   
9.       output;
   
10.    end;
    
11. proc sgplot data = one;
    
12.    histogram mean;
    
13.    density mean;
    
14.    density mean/type = kernel;
    
15. run;

复制代码

Is that what you need?
note: sgplot available in SAS9.2. otherwise, proc univariate also works on the histogram plot.
By the way, I feel like something similar to weak law here. In fact, I am not truely understanding what strong law implies.

2# jingju11

謝謝你回應我,但,可能是因為我用的是SAS9.1,就算不管sgplot還是無法generate date one,在log中出現
169  DM LOG 'CLEAR';
170  DM OUTPUT 'CLEAR';
171  DATA one;
172
173  ARRAY n{20} n1-n20; * a row represents a sample of size 20;
174
175  do sample =1 to 100; *generate 100 samples;
176
177  do n= 1 to 20;
ERROR: The variable type of n is invalid in this context.
ERROR: Illegal reference to the array n.
178
179  n[n] = rannor(1);
ERROR: Illegal reference to the array n.
180
181  end;
182

我真的不知道該怎麼辦!!!

3# sherryshin

1. 
   
2. DATA one;
   
3. array n{20} n1-n20; * a row represents a sample of size 20;
   
4.   do sample = 1 to 100; *generate 100 samples;
   
5.    do i = 1 to 20;
   
6.     n[i] = rannor(1);
   
7.    end;
   
8.    mean = mean(of n1-n20); *100 sample means computed;
   
9.    output;
   
10.   end;
    
11. run;
    
12. ods graphics on;
    
13. proc univariate data = one noprint;
    
14. histogram mean/normal;
    
15. run;
    
16. ods graphics off;

复制代码

Hopefully it works.

4# jingju11
謝謝你,成功了......^O^