Generalized methods for calibration | Michael Rasmussen
| | Abstract | In chemometrics traditional calibration in case of spectral measurements express a quantity of interest (e.g. a concentration) as a linear combination of the sp ectral measurements at a number of wavelengths. Often the spectral measurements are performed at a large number of wavelengths and in this case the number of coefficients in the linear combination is magnitudes larger than the number of observations. Traditional approaches to handling this problem includes Principal Components Regression, (PCR), Partial Least Squares regression, (PLS), Ridge Regre ssion, (RR) and variable selection. They are all presented with theory and appli cation. Least Absolute Shrinkage and Selection Operator, (LASSO), which has recently been improved to handle sigular design matrices, is also presented here. Furthermore a new approach that combines these methods with B-spline basis functions is presented. | | Keywords | Calibration, NIR spectroscopy, linear regression, cross-validation, singular val ue decomposition, regu minimum length least squares, principal components regression, partial least squares regression | | Type | Master's thesis [Academic thesis] | | Year | 2001 | | Publisher | Informatics and Mathematical Modelling, Technical University of Denmark, DTU | | Address | Richard Petersens Plads, Building 321, DK-2800 Kgs. Lyngby | | Series | IMM-EKS-2001-5 | | Electronic version(s) | [pdf] | | BibTeX data | [bibtex] | | IMM Group(s) | Mathematical Statistics |
|