Abstract
This article is focused on approaches used for the detection of purest variables (or samples) in two-way data analysis. The concept of purity is a powerful tool for mixture analysis by spectroscopic or similar methods. Finding rows or columns of the data matrix that carry the contribution from the only component has an independent value helping to better understand the experiment, for instance, to identify mixture constituents or to estimate the quality of chromatographic separation. Detected pure variables can be used as the basis for further “spectral unmixing”—to perform full multivariate curve resolution of the data. Presented methods and algorithms are explained in detail, and their efficiency is illustrated by practical examples.
| Original language | English |
|---|---|
| Title of host publication | Comprehensive Chemometrics : Chemical and Biochemical Data Analysis, Second Edition: Four Volume Set |
| Number of pages | 30 |
| Volume | 2 |
| Publisher | Elsevier Editora |
| Publication date | 1 Jan 2020 |
| Pages | 107-136 |
| ISBN (Electronic) | 9780444641656 |
| DOIs | |
| Publication status | Published - 1 Jan 2020 |
Bibliographical note
Publisher Copyright:© 2020 Elsevier B.V. All rights reserved
Keywords
- Alternating least squares
- Contrast constraint
- Key set factor analysis
- Mixture analysis
- Multivariate curve resolution
- Orthogonal projection approach
- Self-modeling
- SIMPLISMA
- Spectral unmixing
- Spectroscopy
- Step-wise maximum angle calculation
- Stepwise maximum angle calculation
- Variable purity