| Sense Disambiguation Of Ambiguous Words İn Turkish Texts By Machine Learning Algorithms
Word sense disambiguation is the process of selecting the most suitable sense of an ambiguous word in the given sentence by considering the other contextual features
In this study, determination of the most convenient algorithms and features that may lead to the successful disambiguation of the ambiguous words in Turkish texts have been discussed. Ambiguous words and their sense classifications that can be used for Turkish word sense disambiguation studies have been established, a limited ontology has been prepared and by providing manually sense tagged corpora, an important contribution has been achieved for the researches in this domain.
In the first chapters of the thesis, a brief introduction for natural language processing has been given and in parallel to this topic, important work on Turkish have been summarized. Then, the application areas of word sense disambiguation in natural language processing, knowledge types and sources, approaches in the literature and the problems of word sense disambiguation have been examined in the following sections. The objectives, applications and the results of the Senseval project, which is closely related to the thesis topic, have been exhaustively scrutinized
Corpora selection, sense classification, effective features determination, tested algorithms, software and evaluation criterion that have been utilized during the phases of the study have been explained.
The results obtained in the last three years from all phases of the study have been presented by graphics and tables. In the last section, a general evaluation and conclusion have been provided for commenting on the results and future work.
BAHÇETEPE Hürkan ,
Danışman : Prof.Dr. Ahmet SERTBAŞ
Anabilim Dalı : Bilgisayar Mühendisliği
Mezuniyet Yılı : 2006
Tez Savunma Jürisi : Prof.Dr. Ahmet SERTBAŞ (Danışman)
Prof. Dr. Mahmut ÜN
Prof. Dr. Osman Nuri UÇAN
Doç. Dr. Sabri ARIK
Yrd. Doç. Dr. Oğuzhan ÖZTAŞ
Dostları ilə paylaş: | 
