Parameter Redundancy and Identifiability, by Diana Cole

Most quantitative biologists and applied statisticians interested in identifiability-i.e., whether a unique set of parameter values can be found to maximize the likelihood of a model-currently have to swift through piles of manuscripts that are often only tangentially relevant to their needs. A colleague who fits dynamic models to data recently lamented: 'Is there something that I could recommend to my graduate students and postdocs that would not be a very technical book from the 80s or 90s?' Now there is. Diana Cole's new book, Parameter redundancy and identifiability, explores all facets of model identifiability and its little cousin parameter redundancy (when the model is non-identifiable because it could be re-parameterized with fewer parameters). The book is written with mostly ecological modelling in mind (in the broadest sense-including fisheries science, epidemiology, ...). However, as the author notes in the preface, the methodology is general and applies equally to other fields: for instance, to compartmental models from pharmacokinetics, that are covered to some degree, and many other dynamical systems. In many ways, this book builds on the author's previous work (Cole et al., 2010; Cole & McCrea, 2016), but presents identifiability techniques in a more accessible and comprehensive manner than the original papers, which were more focused on so-called exhaustive summaries