HOW TO USE FMATCH (1)

FACTOR MATCHING ALGORITHM BY NORMAN CLIFF, PROGRAM BY PENNELL AND YOUNG
PC-MDS VERSION


FMATCH is a factor matching program that is based on an algorithm by Cliff. The program is designed to orthogonally rotate one or both of two matrices (such as multidimensional scaling solutions) to congruence.

Fmatch will either rotate both of the matrices simultaneously, to a compromise position, or will take one of the two matrices as a target and rotate the other matrix to congruence with the target. The two solution options are known as case I and case II respectively.

Case I, factor matching, simultaneously matches the two solutions is analogous to finding the orientation of r space and orientation of s space and matching the n projections in each of these spaces. The axes of the two spaces are rotated so that the columns of the rotated matrices are as similar as possible. This problem is one of finding eigenroots and eigenvectors and applying those transformations to the two original matrices.

Case II focuses on a fixed target matrix T and attempts to find an orthogonal transformation OT, which, when applied to the original matrix A, produces a matrix A* which is as similar as possible to A. Case II uses the degree of collinearity between corresponding rows of A and A* to measure the congruence of the solution. (The matching procedure is also known as PROCRUSTEAN matching)


FMATCH Parameters:

The data input for FMATCH includes:

PROGRAM LIMITATION: FMATCH can handle a target matrix of up to 50 rows and 50 columns. The data matrix must be of the same dimension. Any number of data matrices can be matched to the target matrix in the same analysis.


FMATCH INPUT:

Line 1: PARAMETERS EXPLANATION (FREE FORMAT)

Line 2: TARGET MATRIX FORMAT

Line 3: DATA MATRIX FORMAT

Line 4+: Supply data as specified above, first the target matrix, followed by the other matrices.

NOTES

1. FMATCH is also known as C-MATCH, or CLIFF MATCH


SAMPLE FMATCH DATA

(The side notes are for explanation and do not belong in the data set):

  10  2  2  1  2
(3X, 2F8.0)               This is input format statement for data set 1
(3X, 2F8.0)               This is input format statement for data set 2
  1 -0.0463  0.5558       This is the start of data set 1
  2  0.5972 -0.0163 
  3  0.5724 -0.3562 
  4 -0.3914 -0.2518 
  5 -0.4079  0.1736 
  6 -0.3826 -0.2096 
  7 -0.4379  0.1296 
  8  0.4792  0.2585 
  9  0.4341 -0.0634 
 10 -0.4169 -0.2203 
  1 -0.3606 -0.4072       This is the start of data set 2
  2  0.4806 -0.2670 
  3  0.5702  0.3010 
  4 -0.3474  0.0876 
  5 -0.4191 -0.1672 
  6 -0.3724  0.1822 
  7 -0.3253  0.3763 
  8  0.5712 -0.1395 
  9  0.5858  0.0547 
 10 -0.3830 -0.0210    


SAMPLE OUTPUT
	F M A T C H 
	FACTOR  MATCHING  PROGRAM 
	WRITTEN BY NORMAN CLIFF 
	PC-MDS  VERSION 
  
 ANALYSIS TITLE: FMATCH TEST DATA
 DATA IS READ FROM FILE: FMATCH.DAT
 OUTPUT FILE IS: B:FMATCH.PRN
  
 INPUT PARAMETERS: 
 DATA IS     10 BY    2 
 TARGET IS   10 BY    2 
 NUMBER OF MATCHES:   1 
 DATA MATCHED TO TARGET AND VICE VERSA 
  
 TARGET 
 ROW: 1    -.046   .556 
 ROW: 2     .597  -.016 
 ROW: 3     .572  -.356 
 ROW: 4    -.391  -.252 
 ROW: 5    -.408   .174 
 ROW: 6    -.383  -.210 
 ROW: 7    -.438   .130 
 ROW: 8     .479   .259 
 ROW: 9     .434  -.063 
 ROW:10    -.417  -.220 
    
 EIGEN ROOTS 
           1           2 
          1.93603      .70970 

 EIGENVECTORS IN COLUMNS 
 ROW: 1    1.000   .000 
 ROW: 2     .000  1.000 
  
 TRANSLATED TO ZERO MEAN AND ROTATED TO PRINCIPAL AXES 
 ROW: 1    -.046   .556 
 ROW: 2     .597  -.016 
 ROW: 3     .572  -.356 
 ROW: 4    -.391  -.252 
 ROW: 5    -.408   .174 
 ROW: 6    -.383  -.210 
 ROW: 7    -.438   .130 
 ROW: 8     .479   .259 
 ROW: 9     .434  -.063 
 ROW:10    -.417  -.220 
   NORMALIZED TARGET 
 ROW: 1    -.091  1.093 
 ROW: 2    1.174  -.032 
 ROW: 3    1.126  -.700 
 ROW: 4    -.770  -.495 
 ROW: 5    -.802   .341 
 ROW: 6    -.752  -.412 
 ROW: 7    -.861   .255 
 ROW: 8     .942   .508 
 ROW: 9     .854  -.125 
 ROW:10    -.820  -.433 
  
 DATA 
 ROW: 1    -.361  -.407 
 ROW: 2     .481  -.267 
 ROW: 3     .570   .301 
 ROW: 4    -.347   .088 
 ROW: 5    -.419  -.167 
 ROW: 6    -.372   .182 
 ROW: 7    -.325   .376 
 ROW: 8     .571  -.140 
 ROW: 9     .586   .055 
 ROW:10    -.383  -.021 

 EIGEN ROOTS 
           1           2 
          2.04309      .56102 
  
 EIGENVECTORS IN COLUMNS 
 ROW: 1    1.000   .000 
 ROW: 2     .000  1.000 

 TRANSLATED TO ZERO MEAN AND ROTATED TO PRINCIPAL AXES 
 ROW: 1    -.361  -.407 
 ROW: 2     .481  -.267 
 ROW: 3     .570   .301 
 ROW: 4    -.347   .088 
 ROW: 5    -.419  -.167 
 ROW: 6    -.372   .182 
 ROW: 7    -.325   .376 
 ROW: 8     .571  -.139 
 ROW: 9     .586   .055 
 ROW:10    -.383  -.021 
  
   NORMALIZED DATA 
 ROW: 1    -.718  -.811 
 ROW: 2     .957  -.532 
 ROW: 3    1.136   .600 
 ROW: 4    -.692   .174 
 ROW: 5    -.835  -.333 
 ROW: 6    -.742   .363 
 ROW: 7    -.648   .750 
 ROW: 8    1.138  -.278 
 ROW: 9    1.167   .109 
 ROW:10    -.763  -.042 
  
 MATCH:    1 
 SIMULTANEOUS TRANSFORMATIONS 
 TRANSFORMATION MATRIX - TARGET TO DATA 
 ORTHOGONALITY OF THIS TRANSFORMATION IS QUESTIONABLE 
  
   1   .998   .067 
   2  -.067   .998 
  
 TRANSFORMED TARGET 
   1  -.164  1.084 
   2  1.174   .046 
   3  1.170  -.624 
   4  -.735  -.545 
   5  -.823   .287 
   6  -.723  -.461 
   7  -.876   .197 
   8   .906   .570 
   9   .860  -.068 
  10  -.789  -.487 
  
 TRANSFORMATION MATRIX - DATA TO TARGET 
 ORTHOGONALITY OF THIS TRANSFORMATION IS QUESTIONABLE 
   1   .996  -.091 
   2  -.091  -.996 
  
 TRANSFORMED DATA MATRIX 
   1  -.641   .873 
   2  1.002   .442 
   3  1.076  -.701 
   4  -.705  -.111 
   5  -.801   .408 
   6  -.772  -.294 
   7  -.714  -.687 
   8  1.159   .173 
   9  1.152  -.215 
  10  -.756   .111 
       *****IDENTIFICATION KEY FOR PLOTS WITH IDENTIFIED POINTS***** 
  
 PT #   1   2   3   4   5   6   7   8   9  10  11  12  13  14  15 
 CHAR   1   2   3   4   5   6   7   8   9   A   B   C   D   E   F 
 PT #  16  17  18  19  20  21  22  23  24  25  26  27  28  29  30 
 CHAR   G   H   I   J   K   L   M   N   O   P   Q   R   S   T   U 
 PT #  31  32  33  34  35  36  37  38  39  40  41  42  43  44  45 
 CHAR   V   W   X   Y   Z   +   /   =   *   &   $   @   %   ?   < 
 PT #  46  47  48  49  50 
 CHAR   (   )   "   #   @    
  
 POINT NUMBERS ABOVE 50 IDENTIFIED AS  >, MULTIPLE POINTS IDENTIFIED AS  #   

 THE FIRST 10 POINTS ARE TARGET---THE LAST 10 ARE DATA 
          +....+....+....+....+....+....+....+....+....+....+....+....+ 
     1.17+                         |                                   + 
         .                      B 1|                                   . 
         .                         |                                   . 
      .96+                         |                                   + 
         .                         |                                   . 
         .                         |                                   . 
      .74+                         |                                   + 
         .                         |                                   . 
         .                         |                        I          . 
      .52+                         |                         8         + 
         .                         |                                   . 
         .     5                   |                                   . 
      .30+    F                    |                                   + 
         .   #                     |                                   . 
         .                         |                                   . 
      .08+-------------------------0-----------------------------------+ 
         .                         |                               C   . 
         .                         |                       J       2   . 
     -.14+                         |                       9           + 
         .                         |                                   . 
         .                         |                                   . 
     -.35+                         |                                   + 
         .    A 6G                 |                                   . 
         .     K4                  |                                   . 
     -.57+      E                  |                                   + 
         .                         |                               D   . 
         .                         |                              3    . 
          +....+....+....+....+....+....+....+....+....+....+....+....+ 
        -1.0  -.8  -.6  -.4  -.2   .0   .1   .3   .5   .7   .9  1.1  1.3 
  
  
  THE GOODNESS OF FIT IS   .889 -1. IS WORST, +1. IS PERFECT  
  
  CORRELATION MATRIX 
 ROW: 1     .972   .000 
 ROW: 2     .000   .654 
  
  THE CORRELATION OF THE DISTANCE VECTORS IS   .856