Custom reparametrization example. The fuller model allows separate learning curves for Task A and Task B, the lesser
model constrains curves to be identical. The mismatch between the empirical and theoretical sampling distribution evident
in the Goodness-of-Fit graph is expected (as these data were generated using an adaptive method). The script PAL_PFLR_CustomDefine_Demo
implements this example and produces a figure like this.
click here to download high-resolution pdf version of picture