Calculates scores for individual units in canonical variates analysis (S.A. Harding).
Option
PRINT = string tokens |
What output to print (scores, adjustments); default scor |
|---|
Parameters
WSSPM = SSPMs |
Within-group sums of squares and products structure |
|---|---|
LRV = LRVs |
Loadings, roots and trace saved from CVA of the WSSPM |
SCORES = matrices |
Unit scores |
ADJUSTMENTS = matrices |
Mean Adjustments |
Description
CVASCORES calculates coordinates of the individual data points projected into the canonical variate space of a canonical variates analysis. This provides data that may be plotted with the canonical variate means to provide more informative graphics.
The WSSPM parameter must be set to the within-group SSP matrix that was used as input to the CVA directive when calculating the analysis, and the LRV parameter must supply the LRV structure formed by CVA. The scores can be saved using the SCORES parameter, and the mean adjustments (as printed by CVA) can be saved using the ADJUSTMENTS parameter. The PRINT option allows the scores and adjustments to be printed, with the default to print just the scores.
Option: PRINT.
Parameters: WSSP, LRV, SCORES, ADJUSTMENTS.
Method
The data matrix X is rotated using the canonical variate loadings, L, to form Y=XL. The columns of Y then have the mean adjustments subtracted, to match the location of the canonical variate means, which are translated such that their their weighted centroid is at the origin. As the procedure needs to compute the mean adjustments, these can also be saved as they are unavailable from CVA itself.
See also
Directive: CVA.
Procedures: CVAPLOT, CVATRELLIS.
Commands for: Multivariate and cluster analysis.
Example
CAPTION 'CVASCORES example'; STYLE=meta
" The data for this example deal with measurements made on 28 brooches
found at the archaeological site of the cemetry at Munsingen. Seven
measurements are used and have been transformed by taking logarithms.
A grouping factor, obtained from a cluster analysis, with four levels
has also been included.
(Doran and Hodson, Mathematics and computers in archaeology. (1975)) "
FACTOR [LEVELS=4] Groupno
READ Groupno,Bow_ht,Bow_thck,Bow_wdth,Coil_dia,Elem_dia,Foot_lth,Length
3 3.21888 3.58352 3.58352 2.83321 2.63906 4.54329 4.74493
1 2.07944 2.89037 3.58352 1.94591 1.09861 3.09104 3.58352
2 2.77259 3.49651 3.68888 2.07944 2.19722 3.52636 4.11087
2 3.29584 4.35671 4.14313 2.30259 2.56495 3.17805 4.31749
2 3.17805 3.97029 4.35671 2.07944 2.19722 3.04452 4.23411
1 2.77259 3.58352 3.63759 2.07944 1.38629 3.3322 4.02535
1 2.94444 3.58352 3.58352 2.3979 1.79176 2.77259 3.71357
4 2.63906 2.70805 5.17615 2.07944 1.79176 3.46574 4.00733
2 2.89037 4.20469 4.5326 1.94591 1.94591 2.99573 3.68888
3 3.17805 4.07754 4.30407 2.48491 2.70805 3.73767 4.27667
3 2.89037 3.8712 4.07754 2.3979 2.19722 3.8712 4.36945
4 2.77259 3.68888 4.77068 1.94591 1.94591 3.4012 3.8712
2 2.63906 3.3322 3.97029 1.94591 2.3979 3.17805 3.73767
2 2.77259 3.89182 3.8712 1.79176 2.56495 3.04452 3.66356
2 2.83321 3.58352 3.95124 2.07944 2.19722 2.89037 3.80666
2 2.77259 3.66356 4.02535 2.07944 1.94591 3.04452 3.93183
2 2.63906 3.80666 3.80666 1.79176 2.3979 3.04452 3.61092
4 2.94444 3.17805 4.40672 2.30259 1.79176 3.09104 3.91202
1 2.89037 3.29584 3.2581 2.3979 2.19722 3.3673 3.98898
3 2.77259 3.78419 3.78419 2.56495 2.48491 4.55388 4.85981
4 2.94444 3.43399 4.48864 2.07944 1.79176 3.13549 4.09434
4 2.70805 2.70805 4.96981 1.94591 1.38629 3.04452 3.80666
2 2.77259 3.85015 3.93183 2.07944 2.63906 3.13549 3.8712
2 3.13549 4.17439 4.23411 2.30259 2.30259 2.56495 3.82864
2 2.77259 3.71357 4.41884 2.3979 2.30259 3.3322 3.98898
1 2.99573 3.58352 3.63759 2.07944 1.38629 2.77259 4.04305
1 2.3979 3.17805 3.04452 1.94591 1.09861 2.3979 3.29584
3 2.94444 4.18965 4.54329 2.30259 1.38629 4.23411 4.70953 :
SSPM [TERMS=Bow_ht,Bow_thck,Bow_wdth,Coil_dia,Elem_dia,Foot_lth,Length;\
GROUPS=Groupno] wssp
FSSP wssp
CVA [PRINT=roots,loadings,means; NROOTS=2] wssp; LRV=lrv; SCORES=mscores
CVASCORES wssp; LRV=lrv; SCORES=uscores