In the literature, various papers have developed more parsimonious Markov model structures concerning full Markov chain models, such as Variable Length Markov Chains (VLMC) and Partition Markov Models (PMM), see [1] and [2]. In the case of a multivariate Markov process, a family of models reduces the number of parameters considering a structure of conditional dependence for the marginal processes (structure of interaction of the process), see [3]. In this work, we infer the interaction structure for three correlates of rhythm for each one of three languages classified as having different rhythmic properties. We show that the structure is different for each language, showing that, this structure can be used as a tool in discriminant computational linguistics.

You do not currently have access to this content.