Fig. 20
From: Approximate parameter inference in systems biology using gradient matching: a comparative evaluation
Distribution of RMS values in function space, calculated on the residuals of the true signal (signal produced with true parameters and no observational error) minus the signal produced with the estimate of the parameters, for the Fitz–Hugh Nagumo model (Eqs. 36, 37). Within each section, the boxplots from left to right are: LB2 method, INF method, LB10 method, GON method and GON method using cross validation for inferring the penalty parameter. The average signal to noise ratio for each “species” is 5. The standard deviation of the observational noise is held at the true value. For definitions, see Tables 2 and 3