Features of "MOG" ("MOrfometria Geometrica")
- Data are coordinates in the FORMAT format (filename_format.txt) after using TET on a TPS file, obtained by either TPSdig or COO.
COO is now able to perform itself the TET conversion into ..._format.txt and into ..._DB.txt at the end of the digitization session.
- The main successive steps of Procrustes superimposition are visualized; different colors are given for different groups if the sample is a subdivided one.
- Partial warps (PW) scores are computed based on total consensus (steps described in the file 'last changes', menu 'version') and saved in a file 'filename_PW.txt'.
- Relative warps (RW) scores are computed as Principal Components of PW scores, and saved in a file 'filename_RW.txt'.
- Possibility is given to assign colors according to groups (Button "SC" for Select Colors).
- Relative landmarks displacements graphically visible as Procrustean coordinates (Button : "Mean Objects"), with the possibility of zooming (Button "ZOOM").
- The MOG module has recent features (Version 91, CLIC34) allowing to
- perform Principal component analysis of residual coordinates ('ALIGNED' specimens), of Procrustes residuals (i.e. differences between residual coordinates and consensus coordinates) and Partial Warps (PW)
- perform Discriminant Analysis on 'Procrustes components' (i.e. principal components of Procrustes residuals) and PW
- enter external, unknown data for classification of unknown individuals if initial data were arranged into 2 or more groups; the classification is based on the shortest distance of each external individual to the average shape of each group and makes use of Procrustes and/or Mahalanobis distances (see the CLIC module information).
- plot and regress shape variables on size (menu "ALLOMETRY") (Version 92, CLIC35)
-
The Mahalanobis classification uses as input the PW (or a few first RW) computed from initial groups plus ONE SINGLE external individual; at each Mahalanobis classification, the PW (and/or RW) are re-computed again... Also note that for the Mahalanobis classification, the discriminant model is computed between initial groups WITHOUT including the external individual (see BMC Research Note for further information).
|
