TNA Fellow Nisa İrem Kırbaç

Dead but Turkish: An NLP Analysis of the Obituaries of the Turkish Community in Belgium

This research project analyses the content of both historical and contemporary obituaries of Turkish people based in Belgium using NLP tools. Specifically, the focus is centered on the way the national identity formation is reflected in these obituaries. The aim is to develop a multilingual (i.e. Turkish, Dutch and French) corpus of death notices from available full-text in newspapers and the Web (e.g. Twitter) and develop a language model for the lesser sourced language of Turkish to conduct NLP analysis on the content of the texts.

Research questions

How can we create a corpus of death notices that belong to Turkishoriginated
migrants in Europe available in newspapers and the Web (e.g. Twitter)?

How can we standardize our corpus to be able to make a meta-data and syntax related
comparison between the results obtained from each source?

How can we use a language model for Turkish, a lesser sourced language, to conduct
NLP analysis on the content of the texts?