% @(#)regrestabl.hlq	17.1.1.1 (ESO-IPG) 01/25/02 17:43:53 
%++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
%.COPYRIGHT  (c)  1990 European Southern Observatory
%.IDENT      regrestabl.hlq
%.AUTHOR     JDP, IPG/ESO
%.KEYWORDS   MIDAS, help files, REGRESSION/TABLE
%.PURPOSE    On-line help file for the command: REGRESSION/TABLE
%.VERSION    1.0  28-OCT-1985 : Creation, JDP
%----------------------------------------------------------------
\se
SECTION./TABL
\es\co
REGRESSION/TABLE					28-OCT-1985  JDP
\oc\su
REGRESSION/TABLE table1 x1[,x2] table2 y1[,y2] degree tol [guess]
	polynomial fit of variables in two tables (not yet implemented)
\us\pu
Purpose:    
            Polynomial fit of one or two variables  between  two tables.
            The command can be used to find the correspondence between
            the entries in the two tables.
            There are two modes of operation:\\
            1. A initial correspondance between the entries is defined
               by the REFERENCE columns in both tables.
               (see command SET/REFERENCE)\\
            2. Initial guesses for the correspondence are given as a
               set of coeffs.\\
            The reference column of the first input table is updated to
            include the references of the correspondent entries in the
            second input table.
\up\sy
Syntax:     
            REGRESSION/TABLE table1 x1[,x2] table2 y1[,y2] degree tol [guess]
\ys\pa
            table1  = first table name
\ap\pa
            x1,[x2] = columns with the values to compare
\ap\pa
            table2  = second table name
\ap\pa
            y1,[y2] = columns with the values to compare
\ap\pa
            degree  = degree of the polynomial used in the
                      transformation (1D or 2D)
\ap\pa
            tol     = tolerance in the correspondance
\ap\pa
            guess   = optional guesses for the transformation in
                      the second mode
\ap\out
Output:
            Output coefficients are listed on terminal and stored in the 
            following keywords:\\
            OUTPUTC   - character information\\
            OUTPUTR(5)- Standard error of the estimate\\
            OUTPUTI(1)- N,no.of data,\\
                   (2)- M,no.of ind.var.\\
                   (4)- col.no. of indep.var.\\
                   (5)- degree (ND)\\
            OUTPUTD(i)- Coefficients\\
                        i=1,...,ND+1
\tuo\no
Note:   
            1.  The main applications are\\
              - In spectroscopy to compute the dispersion coefficients,
                by comparing the user table with line positions and initial
                identifications (table1) with the line catalogue (table2).\\
              - In astrometry to compare object positions from to tables.\\
            2.  The coefficients can be saved (SAVE/REGRESSION) and
                used latter to compute fitted values (COMPUTE/REGRESSION).\\
            3.  The command requires to set the reference columns in both
                tables.
\on\exs
Examples:
\ex
            An application of REGRESSION/TABLE
            The following sequence will identify entries in mytable
            and will compute the dispersion coefficients in a 1D
            application:
\xe\ex
            IDENTIFY/GCURSOR mytable :IDENT :X
\xe\ex
            SET/REFERENCE    mytable :IDENT
\xe\ex
            SET/REFERENCE    lincat  :WAVE
\xe\ex
            REGRESSION/TABLE mytable :X LINCAT :WAVE 5 0.1
\xe\ex
            SAVE/REGRESSION  mytable DISPCOEF
            The command is not yet implemented
\xe \sxe