Extensions to a procedure for generating locally identifiable reparameterisations of unidentifiable systems (Q1595407)

Given a postulated parametrised state space model, structural identifiability is concerned with whether the unknown parameters within the model can be identified uniquely from the experiment considered. A problem that immediately arises is what to do with unidentifiable systems. Some authors have given a methodology for generating locally identifiable reparametrisations of unidentifiable systems. The aim of this paper is to extend such procedures. Symbolic computation, such as Mathematica, facilitates this approach of reparametrisation based on Taylor series. Several concrete examples (biology, pharmacokinetics) are treated. The method generates a reparametrisation that is at least locally identifiable. It is a very interesting paper on models identification.