Table 1. Types of analyses possible with this methodology.

Analysis	# states	# derivatives, rank(${\mathcal{O}}_I$)	other states	parameters	outputs
CCA	s = n	(i = n) OR (ranki+1 = ranki)
PAI	s = n	(i < n) AND (ranki > ranki-1)
PAI	s < n		as unknown p
PAU	s = n	(i = n) OR (ranki+1 = ranki)		removed s.u. p
PAU	s = n	(i = n) OR (ranki+1 = ranki)			o > m

CCA: Complete Case Analysis; PAI: Partial Analysis for Identifiability; PAU: Partial Analysis for Unidentifiability; i: number of Lie derivatives used to build ${\mathcal{O}}_I$; rank_i: rank of ${\mathcal{O}}_I$ with i derivatives; s: number of states taken into account; n: total number of states in the model; o: number of measured outputs; m: number of outputs in the original model; x: states; p: parameters. For detailed explanations, see section “Complete and partial analyses”. In all cases it is possible to remove from the model those parameters that have already been classified as identifiable, see section “Iterative refinement of the diagnosis”.