Using Multiple Sources of Information for Constraint-Based Morphological Disambiguation

USING MULTIPLE SOURCES OF INFORMATION FOR CONSTRAINT BASED MORPHOLOGICAL DISAMBIGUATION G okhan T ur M S in Computer Engineering and Information Science Supervisor Asst Prof Kemal O azer July This thesis presents a constraint based morphological disambiguation approach that is applicable to languages with complex morphology speci cally agglutina tive languages with productive in ectional and derivational morphological phe nomena For morphologically complex languages like Turkish automatic morpho logical disambiguation involves selecting for each token morphological parse s with the right set of in ectional and derivational markers Our system com bines corpus independent hand crafted constraint rules constraint rules that are learned via unsupervised learning from a training corpus and additional statisti cal information obtained from the corpus to be morphologically disambiguated The hand crafted rules are linguistically motivated and tuned to improve pre cision without sacri cing recall In certain respects our approach has been motivated by Brill s recent work but with the observation that his trans formational approach is not directly applicable to languages like Turkish Our approach also uses a novel approach to unknown word processing by employing a secondary morphological processor which recovers any relevant in ectional and derivational information from a lexical item whose root is unknown With this approach well below of the tokens remains as unknown in the texts we have experimented with Our results indicate that by combining these hand crafted statistical and learned information sources we can attain a recall of to with a corresponding precision of to and ambiguity of to parses per token