Statistical programming, matrix languages, and more

Error in calculating std of the residuals in each regression in each simulation

Accepted Solution Solved
Reply
Contributor
Posts: 39
Accepted Solution

Error in calculating std of the residuals in each regression in each simulation

Dear all,

I enclosed a program which simulate regressions

that you can try to run.

I succeeded in calculating the coefficient matrix of the regression (b).

but when I try to calculate the residuals I get just one column  in the matrix that is correct but the other columns are with ones (1).

so the std of the residuals is also is not correct. (bolded with blue colour in the program)

I don't understand why that happens..

Who can help?

Many thanks...

proc iml;

n=10;

b=j(6,n);


     do j=1 to n;                          *Number of simulations;

  resid=j(239,n);

  stdresid=j(n,1);

  xtl = j(240, 1);

  xsl=j(240, 1);

  y=j(240, 1);

  xtl[1,1]=0.926145577;

   xsl[1,1] = 1.367164758;

   y[1,1] = -27.99208409;

  c=0.24269;

  e=0.02106;

  d=0.97564;

  h=-0.24776;

  m=-0.00362;

  l=-0.00111;

  q=0.00091339;

  r=1.00339;

  /** specify the mean and covariance of the population **/

  Mean = {0, 0, 0};

        Cov = {0.0343693345 0.0005718713 -.0000330633,

    0.0005718713 0.0008609739 0.0000083048,

   -.0000330633 0.0000083048 0.0000022582};

  NumSamples = 240;

  call randseed(0);  /** set seed for the RandNormal module **/

  U = RandNormal(NumSamples, Mean, Cov);

            do i=2 to 240;

              Y[i,1]=y[1,1]+u[i,1];         

              xsl[i,1]=c+e*y[(i-1),1]+d*xsl[(i-1),1]+h*xtl[(i-1),1]+u[i,2];    

              xtl[i,1]=m+l*y[(i-1),1]+q*xsl[(i-1),1]+r*xtl[(i-1),1]+u[i,3];    

            end;    

    ylag=lag(y);

    xsl2=xsl##2;

    xtl2=xtl##2;

  

    /* Step 1: Compute X`X and X`y */

    x = j(nrow(y), 1, 1) || ylag || xsl|| xtl || xsl2||xtl2;            /* add intercept column */

  v = {1};                   /* specify rows to exclude             */

    idx = setdif(1:nrow(x), v);  /*  exclude values in v */

    xnew = x[idx, ];             /* extract submatrix                   */

    xpx = xnew`* xnew;                         /* cross-products       */

  v = {1};                   /* specify rows to exclude             */

    idy = setdif(1:nrow(y), v);  /*exclude values in v */

    ynew = y[idx, ];             /* extract submatrix*/

    xpy = xnew` * ynew;

  b[,j]= solve(xpx, xpy);

  resid[,j]=ynew-xnew*b[,j];

  stdresid[j,]=std(resid[,j]);

  end;

  print b;

  print resid;

  print stdresid;

  quit;


Accepted Solutions
Solution
‎10-28-2013 03:48 PM
SAS Super FREQ
Posts: 3,232

Re: Error in calculating std of the residuals in each regression in each simulation

Move the initialization of RESID and STDRESID outside of the DO loop for J.

View solution in original post


All Replies
Solution
‎10-28-2013 03:48 PM
SAS Super FREQ
Posts: 3,232

Re: Error in calculating std of the residuals in each regression in each simulation

Move the initialization of RESID and STDRESID outside of the DO loop for J.

Contributor
Posts: 39

Re: Error in calculating std of the residuals in each regression in each simulation

Rick, thank you!

Post a Question
Discussion Stats
  • 2 replies
  • 263 views
  • 1 like
  • 2 in conversation