A Rule-Based Morphological Disambiguator for Turkish

Part-of-speech (POS) tagging is the process of assigning each word of an input text into an appropriate morphological class. Automatic recognition of parts-of-speech is very important for high level NLP applications, since it would be usually infeasible to perform this task manually in practical systems. One approach to POS tagging uses morphological disambiguation which selects the most suitable morphological parse for each word from the set of parses that is assigned to that word by the morphological analyzer. Accurate POS tagging is not a simple task in general. It even becomes harder for agglutinative languages like Turkish; the number of morphological parses associated with each word in a text is usually much larger than that is for non-agglutinative languages such as English. This is due to the ambiguous nature of such languages. In this paper, we introduce an effective rule based morphological disambiguator for Turkish.