| 摘要 |
A hybrid multivariate spectrographic prediction method combines the qualitative interpretation, spectral shape addition capability of classical least squares multivariate calibration and prediction methods with the facto r analysis methods of partial least squares or principal component regression. The method is used to estimate the quantity of at least one known constituen t or property in a sample through the use of both a classical least squares calibration model and an inverse algorithm such as for example partial least squares or principal components regression.
|
