?===================================================================

?  Condition number for a data matrix.                             =

? User must define the data matrix with the following line:        =

?                                                                  =

?        NAMELIST         ; X = ... $                              =

?===================================================================

? User must change the next line to define data matrix

?

         NAMELIST         ; X = ...$   To define data matrix

?

? XX = X'X.  We then divide X'X(i,j) by the square root of X'X(i,i)*X'X(j,j) to accomplish

? the normalization.  Then, CRX = characteristic roots.  The result would be visible at this   

? point.  CALC just sets up a display. Works for any data matrix.

?

         MATRIX           ; XX  = XDOT(X)                   ?  X'X

                          ; DX  = DIAG(XX)                  ?  Diagonals

                          ; XX  = ISQR(DX) | XX | ISQR(DX)  ?  Normalize

                          ; CRX = ROOT(XX) $                ?  Roots

?

? We report reciprocal to avoid possibly dividing by zero if short rank. 

?

         CALC             ; K = ROW(CRX) ; CN = SQR(CRX(1)) / SQR(CRX(K)) $