Dagger: The Slovak morphological classifier

This paper proposes a classifier, based on hidden Markov model that can be used for solving the problem of part-of-speech tagging of the Slavic languages, such as Slovak, Czech or Polish. These languages are highly inflectional and morphologically rich and have a very large vocabulary. The probability matrices of the classical hidden Markov model are linearly interpolated with additional probability matrices that are calculated using a suffix-based word clustering function. The search space is restricted by a morphological dictionary.