are in very close agreement with the parameter values taken as
true. The largest deviations are observed for the mopt parameter
with difference between 0.3 and 3%. For all other parameters, the
difference with the true parameter is smaller than 2%.
As such, these simulation studies show that the model
parameters can be identified equally accurate from an equidistant
design as from a D-optimal-based design. Whereas in the single
factor case as above, the equidistant level distribution could lead
to identification problems, probably due to an inaccurate situation
of Topt/mopt, this problem does not emerge when identifying
multiple factors. This is most likely due to the combination of data
as a function of T, pH and aw. Following the gamma concept, the
cardinal values related to each of these environmental factors are
independent and, therefore, so is also their parameter identification. However, in the general model, these three environmental
variables are all linked together through the mopt parameter.
are in very close agreement with the parameter values taken as
true. The largest deviations are observed for the mopt parameter
with difference between 0.3 and 3%. For all other parameters, the
difference with the true parameter is smaller than 2%.
As such, these simulation studies show that the model
parameters can be identified equally accurate from an equidistant
design as from a D-optimal-based design. Whereas in the single
factor case as above, the equidistant level distribution could lead
to identification problems, probably due to an inaccurate situation
of Topt/mopt, this problem does not emerge when identifying
multiple factors. This is most likely due to the combination of data
as a function of T, pH and aw. Following the gamma concept, the
cardinal values related to each of these environmental factors are
independent and, therefore, so is also their parameter identification. However, in the general model, these three environmental
variables are all linked together through the mopt parameter.
การแปล กรุณารอสักครู่..
