Model selection in clustering requires (i) to specify a clustering principle and (ii) to decide an appropriate number of clusters depending on the noise level in the data. We advocate an information
