Selecting a set of parameters to be estimated from experimental data is an important problem with many different types of applications. However, the computational effort grows drastically with the number of parameters in the model. This paper proposes a technique that reduces the parameters that need to be considered by clustering, where the model parameters are put into different groups on the basis of the dynamic effect that changes have on the model output. The computational requirements of the parameter set selection problem then drastically reduces as only one parameter per cluster needs to be considered instead of each parameter in the model. This paper develops the underlying theory of the presented technique and also illustrates the method on a model of a Signal transduction pathway with 115 parameters.