Metrohm Vision – Tutorial - Manual (Page 20)

1.3 Developing an Identification method 
NIRS Vision Tutorial – Operation 
13
7. In the Library Method dialog box optionally enter the following
information:
 Comments “Training and today’s date”.
 Click [OK].
Vision now computes the required information on all spectra for all products in
the library. It automatically analyzes each spectrum in the training sets as if it
were an unknown to determine if the method is capable of identifying it.
Figure 2: Results Identification Method
1 Special Outliers 2 Histogram
„Correlation versus frequency“
3 Identification results 4 Percent misidentified versus
product
Review of the results in box 2 (Histogram) shows that all products were
identified with very high correlations. However, in box 3 (samples tab), the
results for stearic acid, calcium stearate and magnesium stearate are listed as
failed.
They are actually ambiguous. Although the samples were identified, one
product had an acceptable match value with another product. This ambiguity
can be handled by using a qualification method to further differentiate
samples. This approach is covered in the next section.