BUCHI NIRCal - Manual (Page 155)

Chemometrics
NIRCal 5.5 Manual, Version A 155
Row Name Description
95 V-Set Regression Slope r= a + bx, where "b"= slope of the V-Set. Should be as close to 1
as possible
Statistics of the original property values (labor method, X-axis)
97 C-Set Orig. min. The smallest property value in the C-Set
98 V-Set Orig. min. The smallest property value in the V-Set
99 C-Set Orig. max. The highest property value in the C-Set
100 V-Set Orig. max. The highest property value in the V-Set
101 C-Set Orig. mean. The mean property value of the C-Set
102 V-Set Orig. mean. The mean property value of the V-Set
103 C-Set Orig. sdev The standard deviation of the property values in the C-Set
104 V-Set Orig. sdev The standard deviation of the property values in the V-Set
106-
113 Statistics of the predicted property values (NIR, Y-axis). These
values should be comparable to the similar values of the original
property values.
115
C-Set RSS C-Set residual sum of squares = RSS should be low:
Sum((Orig-Predicted property)^2) = Sum (e^2)
116 V-Set RSS V-Set residual sum of squares = RSS should be low:
Sum((Orig-Predicted property)^2) = Sum (e^2)
118 C-Set Durbin-Watson dw=Durbin-Watson = Sum of consecutive/succesive Residual
Difference Square / Sum of Residual Square LINK:
Durbin-
Watson Factor Statistics
119 C-Set Durbin-Watson in
range 1.5 to 2.5 Is the dw value between 1.5 and 2.5? Answer: Yes or No
120 V-Set Durbin-Watson dw=Durbin-Watson = Sum of consecutive/succesive Residual
Difference Square / Sum of Residual Square
121 V-Set Durbin-Watson in
range 1.5 to 2.5 Is the dw value between 1.5 and 2.5? Answer: Yes or No
123 C-Set Resid. min Biggest negative deviation in the C-Set property values
124 V-Set Resid. min Biggest negative deviation in the V-Set property values
125 C-Set Resid. max Biggest deviation in the C-Set property values
126 V-Set Resid. max Biggest deviation in the V-Set property values
128 V-Set t-value Original property and the Predicted property are compared via a
paired t-test to show if the results are statistically equivalent
129 V-Set t-value (n-1, 2-tail)
Confidence (%) Significance level, where is no evidence for a difference between
the Original and Predicted property results
131 C-Set n Number of C-Set spectra
132 V-Set n Number of V-Set spectra