Graduate Program in Cognitive Science.Güngör, Tunga.Atlamaz, Ümit.Dursun, Olgun.2025-04-142025-04-142023Graduate Program in Cognitive Science. BM 2023 K38 PhD (Thes CHEM 2023 E78https://digitalarchive.library.bogazici.edu.tr/handle/123456789/21648Morphological parsing is the computational task of breaking down words into their roots and affixes. There are several successful morphological parsers for Turkish, especially for inflectional morphology. However, there is a gap in the literature concerning the analysis of fusional properties of foreign-origin words, support for prefixes, and comprehensive derivational suffix coverage. To address this gap, this thesis describes and implements a new computational morphological processing framework for Turkish with novel principles. These principles are based on the recent opportunities and requirements in the natural language processing field, namely the transformer-based pre-trained large language models and fine-tuning approaches. The framework contains a description of language resources structure, a morphological analyzer that examines all possible parses of a word, a morphological disambiguator that picks the correct hypothesis among analyzer outputs, and error analysis modules for these tools.Turkish language -- Morphology.Morphology & Syntax.ix, 59 leaves