CHEMOMETRICS ASSISTED UV-SPECTROPHOTOMETRIC ...

A. Advantages of PCR
• PCR doesn’t need the selection of wavelength most of the time the whole spectrum
are used
• Averaging effect: as one uses great number of wavelengths the averaging effect will
be attained decreasing the chance for spectral noise can be utilized for mixtures with
large constituents (highly complex). PCR also enables, some times to figure out
samples with constituents which are not present basically (originally) in the calibration
mixture.
B. Disadvantage of PCR
• The calculation is slower if compared to CLS.
• Optimization needs knowledge of PCA i.e. interpretation and understanding the model
is not a simple task.
• It needs large number of samples for the accurate calibration [32,37].
Insipite of the above inconveniences, PCR has widely been applied for the specrophtometric
resolution of mixtures comprising two or more serous overlapping spectra [38,39].
1.3.3. Partial least squares
PLS is one of the factor analysis methods that widely used and gives complex information
[12, 40]. It is a powerful multivariate statistical tool applied to simultaneous
spectrophotometeric determination of mixture with overlapping spectra (of constituents). PLS
is the method that is often used in multivariate calibration. It resembles PCR, but works in
one step. It also determines latent variables, that are linear combination of the original
variables, but the criterion applied is maximal covariance between the Y-values and the
spectral variables because of this criterion the algorithm yields models by and interactive
procedure which is perceived by the user as a single regression step [39].
A. Advantage of PLS
• It works in one step unlike the PCR
• Can be used for very complex mixture
9