 Usage:  (ra,b,q)=hurst(x,k)  
 
 Input:

  x                      n x 1 vector, observations of the process 
                         
  k                      scalar, maximal number of intervals for the R/S statistic 
                         
 Output:

  ra                     (k-2) x 3 matrix, ra[,1] = (1, ... ,1)', 
                         ra[,2] = log(n/3, n/4, ... ,n/k)', 
                         ra[,3] = log(RS) 
                         
  b                      2 x 1 vector, b[1] = intercept of the R/S-line, 
                         b[2] = slope of R/S line, 
                         b solves the 
                         regression problem ra[,3] = b[1] + b[2]*ra[,2] 
                         
  q                      scalar, residual varince of the regression problem 
                         
--------------------------------------------------------------
(C) MD*TECH Method and Data Technologies, 17.8.2000
