Metrohm Vision – Theory Instrukcja Użytkownika

NIRS Vision – Theory Manual 8.105.8011EN

8 ▪▪▪▪▪▪▪ equal to zero (0). Formally R² indicates the fraction of total variance in the data set modeled by the equation. 1.4.2 Standard Error of C

▪▪▪▪▪▪▪ 9 2 Math Pretreatment Methods 2.1 N-Point Smooth This is a boxcar type of smoothing. The method’s parameter, segment size, defines the size

10 ▪▪▪▪▪▪▪ value computed as B-A is assigned to the data point in the middle of the gap. Then the whole segment-gap-segment sequence is shifted one d

▪▪▪▪▪▪▪ 11 derivative spectra will exhibit offset variation. It is common practice, therefore, to take the second derivative with respect to w (wave

12 ▪▪▪▪▪▪▪ Specified Segment Size Points per segment, s Specified Gap Size Points per gap, g 1 – 2 1 1 – 2 1 3 – 6 3 3 – 6 3 7 – 10 5

▪▪▪▪▪▪▪ 13 The polynomial function models the effects in a cumulative fashion as its order increases from 0th to 1st and 2nd degrees: Order of the Po

14 ▪▪▪▪▪▪▪ linear regression is performed on absorbance values of the sample spectrum versus those at corresponding wavelengths in the mean spectrum.

▪▪▪▪▪▪▪ 15 3 Qualitative Library Development 3.1 Principal Component Analysis The near IR spectrum comprises intensity measurements at hundreds of

16 ▪▪▪▪▪▪▪ Multiplication of a spectrum by a set of eigenvectors yields a set of scores, which can be interpreted as coordinates of the spectrum in t

▪▪▪▪▪▪▪ 17 3.2.3 Maximum Distance To calculate maximum distance (distance in this context is in terms of spectral value – absorbance, or the spectra

18 ▪▪▪▪▪▪▪ 4 Sample Selection Methods 4.1 Mahalanobis Distance in Principal Component Space 4.1.1 Outlier Detection Principal Component Analysis p

▪▪▪▪▪▪▪ 19 estimates the what threshold is required to obtain the desired distribution of samples between training and acceptance sets. Before the sa

20 ▪▪▪▪▪▪▪ 4.4 Sample Selection Based on Lab Data (Quantitative) In this method of sample selection Vision displays a histogram of lab data distribu

▪▪▪▪▪▪▪ 21 5 Identification and Qualification Methods 5.1 Wavelength Correlation 5.1.1 Model Development The first step in development of a wavele

22 ▪▪▪▪▪▪▪ Another type of threshold is match value. Mahalanobis distance calculated directly from the formula depends strongly on the number of samp

▪▪▪▪▪▪▪ 23 5.4.2 Analysis of an Unknown Using a products’ principal component model, the residual variance of the unknown spectrum is calculated usi

24 ▪▪▪▪▪▪▪ 6 Library Clustering 6.1 General Description Library clustering is an algorithm designed to generate a simplified representation of a la

▪▪▪▪▪▪▪ 25 Once these parameters have been defined, the algorithm proceeds as follows: 1. A principal component model is calculated for the library

Metrohm AG CH-9100 Herisau Switzerland Phone +41 71 353 85 85 Fax +41 71 353 89 01 [email protected] www.metrohm.com NIRS Vision – Theory

Teachware Metrohm AG CH-9100 Herisau [email protected] This documentation is protected by copyright. All rights reserved. Although all

▪▪▪▪▪▪▪ 3 Table of contents 1 Calibration Development ...

4 ▪▪▪▪▪▪▪ 4 Sample Selection Methods ... 18 4.1 Mahalanobis

▪▪▪▪▪▪▪ 5 1 Calibration Development 1.1 Simple Linear Regression 1.1.1 Overview The simplest method of calibration is based on a single independen

6 ▪▪▪▪▪▪▪ 1.2 Multilinear Regression 1.2.1 Overview Multilinear Regression (MLR) is an extension of the simple linear regression. This method uses

▪▪▪▪▪▪▪ 7 1.3.2 Preprocessing Calibration Data Spectral and constituent data are routinely preprocessed before the PLS calibration. (Here, the term