Key points are not available for this paper at this time.
A binary FSMX source generates a given sequence of digits from 0, l whose statistical behavior can be described using a postfix set S. This postfix set is a collection of binary strings which is proper and complete. To weigh an infinite number of models, the authors introduce a principle which says that the model-redundancy has to be proportional to the number of free parameters of the model. It gives us a weighting distribution over all models. A efficient method is outlines that weighs the block probabilities of all models according to this distribution.
Willems et al. (2005) studied this question.