Recognition
Classify and analyze your new spectra of unknown samples at a mouse click. If you have defined more than one
class assignment scheme
in order to answer several different questions on the basis of your spectra, you may perform all the classifications with one click.
As a result, you obtain the system's class decisions for the individual spectra - along with corresponding confidence values for the
prediction correctness - presented in a table. In addition, you may check for some given spectrum the (lower) confidence values
for those classes, which have not been assigned by the system, and see whether some different class decision could be considered.
Go through the results for the individual spectra with a fine-tooth comb. Compare the unknown spectra with reference spectra from your
Knowledge Base and recheck the class decisions.
In addition, you may employ the so-called "use flags" to define some of your known spectra to be "unused", i.e. excluded from all training steps.
In this way, the unused spectra remain completely unknown to the system. After having set up and trained a classifcation system,
you may use it to classify the unused spectra and compare the results with the known actual class assignments. This is yet another method of
evaluation for the classification system.
Spectroscopy Techniques 
