看了前人的帖子，已经安装别人说的，相应的文件粘贴到lib以及src里了
程序如下：
$ n=31;
$ t=20;
$ k=1;
$ load s[t,n]=C:\gauss9.0\ex519\y.txt;
$ load z[t,n*k]=C:\gauss9.0\ex519\fli.txt;
$ library panel coint;
$ _ker_fun=&fejer;
$ output file=result.out reset;
$  dols_rd(s,z,"result");

(0) : error G0159 : Wrong number of parameters
Currently active call: DOLS_RD
Stack trace:
   DOLS_RD
再者，用fmols
$ n=31;
$ t=20;
$ k=1;
$ load s[t,n]=C:\gauss9.0\ex519\y.txt;
$ load z[t,n*k]=C:\gauss9.0\ex519\fli.txt;
$ library panel coint;
$ _ker_fun=&fejer;
$ output file=result.out reset;
$  fm_rd(s,z,"result");
The Conventional T test and OLS estimator
The FM_beta1 is:  0.2866
The t-ratio1 is: 17.0631 probabiltiy(t)=  0.0000 probability(N)= 0.0000
R square is:  1.3786
Adjusted Rsquare is: -0.1363
The conventional Covariance Matrix:
0.0006  0.0000
0.0000  0.0020
The autocorrelation Matrix is:
0.0006  0.0001
0.0003  0.0008
The Long-run Covariance Matrix is:
0.0019  0.0004
0.0004  0.0035
The Inverse of Long-run Covariance is:
542.3879 -64.1959
-64.1959 293.0632

The true Covariance of FM_beta is:  3.1579
这个结果不对，应该不是fmols
原来一直用stata，要做非平稳面板估计，刚入手gauss，看了以前的帖子，问题都是讨论一般就中断了。有没有会做的童鞋能指点一下