Metrohm Vision – Tutorial - Manual (Page 32)

1.7 Partial Least Squares Regression 
NIRS Vision Tutorial – Operation 
25
 Training Set 75 %
 Acceptance Set 25 %
 Click on [OK]
Vision now performs calculations according to the parameters defined in the
previous steps.
The Results of the Sample Selection are displayed (cf. Figure 5, page 20).Figure
5
Result Sample Selection for quantitative Analysis
Save Results
1. Click on Select  Save Results.
2. If you are asked “Warning – By saving this calibration equation you
will invalidate the regression method which you previously
developed for this equation. You must develop the regression again
to use this equation. Do you want to continue?”, click on [YES].
3. If you are asked “Do you want to save redundant samples as
Validation Set?” Click on [YES].
1.7.2 Regression
Configure Regression Method
1. Click on Mode  Quantitative Analysis  Regression.
2. Double-click on the xdsstart project name.
3. Double-click on the product “HiPro”
the tree expands to show all the products and their associated samples,
constituents, and equations when available.
4. Double-click on (f(x)) Equations under “HiPro”.
5. Double-click on f(x) Protein
the name given for the equation turns light blue indicating that it has
been selected for Calibration Development.
6. Click on Regression  Edit Regression Method.
7. In the Edit Regression Method dialog box, enter the following:
 Regression Partial Least Square
 Pre-Treatment Method 2nd derivative
 Leave all other settings unchanged