Specifies supported objective function configurations for error calculation
in parameter optimization. These specifications detail the allowable options
for tailoring how model fit is assessed, configured within PIConfiguration
.
Details
Includes:
objectiveFunctionType
– Type of error calculation. Allowed values:lsq
(least squares),m3"
(M3 method).residualWeightingMethod
– Method for weighting residuals. Allowed values:none
,std
,mean
,error
.robustMethod
– Approach for robust outlier handling. Allowed values:none
,huber
,bisquare
.scaleVar
– Boolean for variance scaling. Allowed values:TRUE
,FALSE
.scaling
– Scaling type. Allowed values:lin
(linear),log
(logarithmic).linScaleCV
– Coefficient of variation for linear scaling. Numeric range:1e-9
to1
.logScaleSD
– Standard deviation for log scaling. Numeric range:1e-9
toInf
.
These options directly influence the optimization process by defining how discrepancies between simulated and observed data are quantified and managed.