Quantifying Knowledge Production: NLP-Driven Insights and Forecasting Models for Thesis Output in Turkey’s Basic Sciences

Authors : Şeyma Bozkurt Uzan, Nesrin Manav, Şevval Arabacıoğlu

Pages : 287-300

Doi:10.7240/jeps.1673780

View : 131 | Download : 231

Publication Date : 2025-12-23

Article Type : Research Paper

Abstract :In the contemporary world, research in the field of basic sciences is predominantly conducted within academic institutions. Graduate theses play a pivotal role in this process, as they not only represent the initial steps on an individual\\\'s academic journey but also their contributions to scientific knowledge. This study presents a comprehensive bibliometric analysis of graduate theses written in Türkiye in the basic sciences from 1960 to 2024. The research aims to investigate the trends in the number of theses over time, their distribution by years, types, and universities, as well as visualizing recurring keywords and research themes through a word cloud. By using the Instant Data Scraper program to gather data from the National Thesis Center, a total of 68,883 theses were analyzed. The statistical methods employed include descriptive statistics, time-series forecasting, and natural language processing (NLP) techniques. The results indicate a steady increase in thesis production over the years, with significant contributions from Gazi University and Middle East Technical University (METU). While chemistry and biology show strong activity, mathematics and theoretical sciences are underrepresented. Forecasting models predict continued growth in thesis production, with variations reflecting external influences. These results provide valuable insights into Türkiye’s graduate research landscape and can guide future academic and policy development.
Keywords : Doğal Dil İşleme (NLP), Zaman Serisi Tahmini, Lisansüstü Tezler, Bibliyometrik Analiz, Türkiye`de Temel Bilimler

ORIGINAL ARTICLE URL