On the similarity between two approaches to the identifiability problem (Q1099119)

Two methods for studying the identifiability of linear models are particularly analysed: The ``modal matrix approach'' and the ``exhaustive modeling approach'', the first one proposed by the author [Math. Biosci. 52, 73-96 (1980; Zbl 0455.93019); ibid. 70, 1-37 (1984; Zbl 0562.92001)] and by \textit{J. P. Norton} [ibid. 50, 95-115 (1980; Zbl 0433.93064)], and the second by \textit{S. Vajda} [Int. J. Syst. Sci. 14, 1229-1247 (1983; Zbl 0527.93018)]. These strategies both operate in the eigenvector-space and, although apparently quite different, they show, in the light of a careful analysis, to be similar methods. After a brief summary of the two approaches the author compares step by step the equations involved and points out that, although by utilising different unknowns, they deal with the same properties and the general results obtained by one approach can also be reached by the other.