Principal component regression in NIR analysis: Viewpoints, background details and selection of components
Tóm tắt
In this paper we present formulae for prediction error related to principal component regression (PCR). The difference between PCR and ordinary least‐squares (LS) regression is discussed in relation to these formulae. This discussion is used as a basis for a treatment of PCR in NIR analysis. The theory is illustrated by two examples from NIR analysis.
Từ khóa
Tài liệu tham khảo
Osborne B. G., 1986, Near Infrared Spectroscopy in Food Analysis
Martens H., 1987, Near‐infrared technology in the agricultural and food industries, 57
Weisberg S., 1985, Applied Linear Regression
Mittelhammer R. C., 1977, Amer. J. Argic. Econ., 336
Mardia K. V., 1979, Multivariate Analysis
Wold H., 1966, Multivariate Analysis
Stone M., 1973, J. Roy. Statis. Soc., B, 36, 111
T.Næs ‘Leverage and influence measures related to PCR’ Report Norwegian Food Research Institute (1986).