A Socio-Political Network Analysis from the Late Ottoman Empire to Early Republican Periods: Re-evaluating Personal Narratives through Artificial Intelligence (1900-1940)
Re-interpreting history by analyzing period testimonies from 1900-1940 with artificial intelligence
This project is designed to understand the impact of interpersonal relationships and socio-political networks on political history during the final period of the Ottoman Empire and from the establishment of the Republic to the 1940s, utilizing artificial intelligence technology.
Using 210 sources witnessing the period between 1900-1940, socio-political network analysis is being conducted through texts containing period testimonies such as memoirs, reminiscences, autobiographies, and diaries.
The project focuses on the emotional states between individuals at word, sentence, and paragraph levels through natural language processing (NLP) tools to understand the dimensions of interpersonal relationships.
Memoirs, reminiscences, autobiographies and diaries containing period testimonies
Modern language models and sentiment analysis techniques
Mapping and visualization of interpersonal relationships
Analysis of historical data with an interdisciplinary approach
Our academic experts in history, linguistics and computer science
Project Director
PhD: Boğaziçi University, Institute of Atatürk Principles and Revolution History, 2021
Researcher
PhD: Dokuz Eylül University, Institute of Atatürk Principles and Revolution History, 2010
Researcher
PhD: Dokuz Eylül University, Turkish Language and Literature, 2012
Researcher
PhD: Kocaeli University, Turkish Language and Literature, 2021
Consultant
PhD: Ege University, Computer Engineering, 2010
Researcher
PhD: Dokuz Eylül University, Computer Engineering, 2021
MA Student
BA: Pamukkale University
MA Student
BA: Dokuz Eylül University
PhD Student
BA & MA: Dokuz Eylül University
Review our models and research outputs published on HuggingFace and GitHub
A sentiment analysis model developed to analyze general emotional states in historical texts from the late 19th and early 20th centuries. It detects positive, negative and neutral emotional states with high accuracy.
A sentiment analysis model developed to analyze interpersonal relationships in historical texts. Our model developed with the DECA-EBSA method is designed to understand the author's emotional state towards individuals in sentences.
NER model that automatically identifies person, place and institution names in historical texts.
Application tool for digitizing complex and irregular texts from the 19th and 20th centuries and Optical Character Recognition.
Interactive network analysis tools developed to visualize interpersonal relationships. Offers rich visualizations with Gephi, NetworkX and D3.js integration.
A labeled and structured dataset consisting of period testimonies from 1900-1940. It is offered as open access for researchers.
Izmir Institute of Technology
General Culture Department
Gülbahçe, Urla, Izmir 35430
You can contact us to get detailed information about the project, collaborate or share your research questions.
Send Message