Probabilistic, Information-Theoretic Models for Etymological Alignment
Hannes Wettig · Työväentutkimus Vuosikirja · 2013
This thesis starts out by reviewing Bayesian reasoning and Bayesian net-work models. We present results related to discriminative learning of Bayesian network parameters. Along the way, we explicitly identify a num-ber of problems arising in Bayesian model class selection. This leads us to information theory and, more specifically, the minimum description length (MDL) principle. We look at its theoretic foundations and practical impli-cations. The MDL approach provides elegant solutions for the problem of model class selection and enables us to objectively compare any set of mod-els, regardless of their parametric structure. Finally, we apply these meth-ods to problems arising in computational etymology. We develop model families for the task of sound-by-sound alignment across kindred languages.