Machine learning using pseudolandmarks

How to use the MACHINE LEARNING section of XYOM for outlines data ?

  • Do not enter directly the raw coordinates of the pseudolandmarks
  • Compute and save the NEF from unknown data (or PCs of NEF)                             <– CHARACTERIZATION
  • Compute and save the NEF from reference data (or PCs of NEF)                             <– CHARACTERIZATION
  • Select the same number of columns for each file (unknown and reference)        <– MISCELLANEOUS (Working with data files, Extraction)
  • Enter these data in the Multilayer Perceptron, enter the subdivision of the reference file, activate the SUBMIT button and see the report

 

If WEIGHTS were saved, new unknown (transformed to NEF or PCs of NEFs, with the same number of columns used for the learning process) could be tentatively identified using the IDENTIFICATION section of XYOM.


 

Another way to apply ANN to outlines data is possible, as follows

  • Do not enter directly the raw coordinates of the pseudolandmarks
  • Concatenate the raw pseudolandmarks of unknown and reference data                  <– MISCELLANEOUS (Working with data files, Concatenate, By Rows)
  • Compute the NEF (or PCs of NEF) from the concatenated unknown and reference data <– CHARACTERIZATION
  • Extract and save the unknown NEF (or PCs of NEF)                                              <– MISCELLANEOUS (Working with data files, Extraction)
  • Extract and save the reference NEF (or PCs of NEF)                                              <– MISCELLANEOUS (Working with data files, Extraction)
  • Enter these data in the Multilayer Perceptron, enter the subdivision of the reference file, activate the SUBMIT button and see the report

 

If weights were saved, new unknown (transformed to NEF or PCs of NEFs, with the same number of columns used for the learning process) could be tentatively identified using the IDENTIFICATION section of XYOM.