Diploma Thesis DIP-2017-05

BibliographyKöstak, Ugur: Language Identification for German-Turkish Code-Switching Speech.
University of Stuttgart, Faculty of Computer Science, Electrical Engineering, and Information Technology, Diploma Thesis (2017).
69 pages, english.
CR-SchemaI.2.7 (Natural Language Processing)
Abstract

The importance of computers has risen in recent years in our daily lives. An average person interacts without a doubt multiple times with computers. The wide usage of computers has caused researchers to think of ways which would allow you to communicate with computers by a minimum number of interactions. Speech is the main communication instrument for humans, so researchers also used speech as an interaction method between humans and computers. However, speech has boundaries of its own, the language varies in different societies, especially in multicultural societies where people tend to use a mixed language called Code-Switching language to communicate, i.e. Germany is a multicultural country and foreigners, especially bilingual Turkish people, use German and Turkish when they speak to each other. On the other hand, computers nowadays have become more powerful and can also process complex tasks such as NLP tasks, which requires a lot of processing power. In this thesis we aimed to solve Language Identification task in German-Turkish code-switching speeches with two popular machine learning methods Support Vector Machines and Deep Neural Networks and at the end we compared the performances of these methods.

Full text and
other links
PDF (1489082 Bytes)
Department(s)University of Stuttgart, Institute for Natural Language Processing
Superviser(s)Vu, Jun.-Prof. Ngoc Thang; Schweitzer, Dr. Antje; Cetinoglu, Dr. Özlem
Entry dateJuly 3, 2018
   Publ. Computer Science