CCL Home Page
Up Directory CCL README
This program was retrieved from anonymous ftp on

This programs fits a function (given as a FORTRAN expression) to data.
Read fitinstr.txt for more information.

Before you can try to use it on your system, you need to change all accurances
of /usr/local/src/splx to the path where the source files for this
program really reside.

You need a Freelib.a to link this program. To make this library
get stuff from freeread directory in software/SOURCES/FORTRAN and run make
on it.

Jan Labanowski


          The  fit  utility  will  generate,  compile,  link,  and  execute  a
    FORTRAN  program to fit a set  of  data points to a user-provided equation
    using the SIMPLEX algorithm.   The data points and initial guesses for the
    adjustable  parameters  will  be read from a file.   The  equation for the
    function  has  to be provided when prompted as a single,  correct  FORTRAN
    statement evaluating a variable called YCALC, e.g. :

                    YCALC = A(1) + A(2)*X(1) + A(3)*X(2)

    where the A(i)'s represent the parameters to be determined, and the X(i)'s
    the  independent variables.   The equation can be non-linear in either the
    parameters  or the variables.   Both the variables and the parameters have
    to  be given in the same sequence that they appear in the input file,  and
    the  indeces  have  to  be  continuously  increasing, unless refering to a
    previously defined variable or parameter, i.e. :

      the equation YCALC=A(1)+A(2)*X(1)+A(4)*X(2)       is NOT valid since the
                                                    sequence of A's is broken.

      the sequence YCALC=A(1)+A(2)*X(1)+(A(1)+A(2))*X(2)             is valid.

    the  input  equation  is  terminated by a , thus a single line must be
    used.  This line can however be up to 256 characters long and wraps around
    when the edge of the screen is reached.  The equation is parsed, converted
    to lower case, and  broken  into  all  the necessary FORTRAN lines by this

       The input file should contain in the first line the initial guesses for
    the  values  of  all  the  parameters  in  the sequence established by the
    indeces assigned in the equation.   The remaining lines should contain the
    values  for  all  the  variables  in  the  sequence  of their assignement,
    followed by the "observed" value of the dependent variable (Y).  All these
    values are read in free format and the program will recognize normal real,
    exponential,  or integer numbers as input.  The only requirements are that
    all  the  values  for the independent and dependent variables are given in
    the  same line for each point, and that they are separated by at least one
    blank.   Alphanumeric text or comments can be present and the program will
    ignore them.
Modified: Mon Dec 25 17:00:00 1995 GMT
Page accessed 4820 times since Sat Apr 17 22:01:37 1999 GMT