请教应用sas如何进行蒙特卡罗模拟

shiyiming

我要求获得4个变量，1000组不相关、独立的服从N（0，1）分布的数据集，请问用sas如何实现？

shiyiming

[quote="gbt":a99aa]我要求获得4个变量，1000组不相关、独立的服从N（0，1）分布的数据集，请问用sas如何实现？[/quote:a99aa]

[code:a99aa]data a;
do i=1 to 1000;
  x1=rannor(1);
  x2=rannor(2);
  x3=rannor(3);
  x4=rannor(4);
  output;
end;
run;[/code:a99aa]

Please be noted that you are talking about independence between x_i and x_j, not records.

shiyiming

生成的是随机数字吗？为什么每次运行的结果都是一样？

shiyiming

it's pseudo-random numbers, fixed when the seed is fixed except you use 0 as the seed.

shiyiming

Sorry, it should be written as

[code:d786f]data a;
do i=1 to 1000;
  x1=rannor(0);
  x2=rannor(0);
  x3=rannor(0);
  x4=rannor(0);
  output;
end;
run;[/code:d786f]

in case you do not want to see the repeated measure.

shiyiming

rannor(0)
参数为0时,随机数种子为系统时间,所以每次不一样,如果你不为0,每次当然是
同一组随机数了(随机数种子相同)