Last Update: 2/4/2015
A global fit will fit all datasets simultaneously, allowing parameter sharing between the fit curves of each dataset. An independent fit will fit each dataset one at a time.
Fitting datasets with the global setting (with no parameter sharing) as compared with the independent setting will typically result in different parameter errors, and possibly different parameter values for the fit. This is because global fitting performs minimization in a combined parameter space.
Global fitting is appropriate/necessary only if you want to share parameters between the fit curves of the datasets.
Keywords:nonlinear, regression, multiple datasets