Assessment of the structure and predictive ability of models developed formonitoring key analytes in a submerged fungal bioprocess using near-infrared spectroscopy
S. Vaidyanathan et al., Assessment of the structure and predictive ability of models developed formonitoring key analytes in a submerged fungal bioprocess using near-infrared spectroscopy, APPL SPECTR, 55(4), 2001, pp. 444-453
The robustness of models developed for the near-infrared spectroscopic pred
iction of mycelial biomass, total sugars, and ammonium in a submerged Penic
illium chrysogenum bioprocess was assessed by rigorously challenging them w
ith artificially introduced analyte and background matrix variations, so th
at analyte concentrations were varied in an invariant matrix and vice versa
, The models were also challenged by using a data set from a process operat
ed at a different scale from that used in the original model formulation. S
imple univariate and bivariate linear regression models, and partial least-
squares (PLS) models with as few; factors as three and four, performed suff
iciently well for predicting analyte concentrations and were robust with re
spect to the matrix variations tested, Howe c er, models based on relativel
y weaker absorptions, or those that were likely to be influenced by stronge
r absorbers present in the same matrix, were vulnerable to changes in the m
atrix, ri change in the stale of operation affected models that would he in
fluenced by biomass, possibly due to an influence of tile morphology of the
mycelial biomass. An analysis of the loading vectors of some PLS models re
vealed details that were useful in understanding the type of information mo
deled and the behavior of these models to the variations tested.