Türçe varoluşsal cümlelerin destek vektör makineleri kullanılarak anabilimsel argüman sınıflandırılması ve anlambilimsel gruplanması

ABSTRACTSEMANTIC ARGUMENT CLASSIFICATION ANDSEMANTIC CATEGORIZATION OFTURKISH EXISTENTIAL SENTENCES USINGSUPPORT VECTOR LEARNINGAylin KocaM.S. in Computer EngineeringSupervisor: Prof. Dr. Varol AkmanSeptember, 2004There are three types of sentences that form all existing natural languages: verbalsentences (e.g. ?I read the book.?), copulative sentences (e.g. ?The book is on thetable.?), and existential sentences (e.g. ?There is a book on the table.?). Syntactic andsemantic recognition of these sentence types are crucially important in computationallinguistics although there has not been any significant work towards this end. Thisthesis, in an attempt to fill this evident gap, is on identifying and assigning semanticcategories of Turkish existential sentences in print. Existential sentences in Turkish areminimally characterized by the two existential particles var, meaning there is/are, andyok, meaning there is/are no. In addition to these most basic meanings, other senses ofexistential particles are possible, which can be categorized into groups such as caseexistentials and possession existentials. Our system does shallow semantic parsing indefining the predicate-argument relationships in an existential sentence on a word-by-word basis, via utilizing Support Vector Machines, after which it proceeds with thesemantic categorization of the whole sentence. For both of these tasks, our systemproduces promising results, in terms of accuracy and precision/recall, respectively. Partof this research contributes to the annotation of the METU-Sabancı Turkish Treebankwith semantic information.Keywords: shallow semantic parsing, semantic role labeling, thematic roles, supportvector machines, Turkish existential sentences, Turkish Treebank.iii