ABSTRACT
This Chapter explores the problems associated with detecting parameter redundancy and identifiability in complex models. As models get more complex, the algebra involved in the symbolic methods also becomes more complex. Symbolic algebra packages, such as Maple, can run out of memory trying to execute this method. Extensions for the symbolic algebra method are discussed, including methods for models with covariates and a more general extension, using reparameterisation, which is suitable for most models.
The alternative is to use a hybrid of symbolic and numerical methods. This Chapter also discusses extensions to this hybrid-symbolic numerical method.
