Akademik Veri Yönetim Sistemi
Current Issues in Archival Science, İshak Keskin,Ceyhan Güler,Sinan Sümbül, Editör, Istanbul University Press, İstanbul, ss.90-104, 2025
Probabilistic subject models, which are expressed as a group of algorithms and are revealed by transforming the
hidden thematic information in the relevant and related documents or resources in a certain field into a small-sized
universe, are among the research topics that are used in the fields of machine learning and text mining and whose
importance continues to increase. Machine learning and text mining, which are also used in the field of archival
and records management, provide benefits for the field, especially in terms of identifying topics in large volumes
of resources and visibility to missing or unstudied topics. Within the scope of this study, applying Latent Dirichlet
Allocation (LDA), which is a subject modelling method that performs automatic and semantic analysis of master's and
doctoral theses on archival and records management prepared in Information and Records Management Departments
(BBY) in Türkiye. In the LDA application, the experimental study on the extracts of 243 theses prepared and obtained
between 1987-2022 on archival and records management issues included in the National Thesis Center Database of
the Council of Higher Education and in the thesis catalogues of the universities where BBY departments are located,
was carried out with the Gibbs sampling algorithm. As a result of the study, 20 main topics were determined. Namely
"corporate electronic records management systems", "electronic records management", "television archives management", "Ottoman diplomacy", "ERMS and access services", "management of visual archives", "artificial intelligence and
access", "digital cultural heritage management”, “web accessibility of cultural heritage”, “corporate electronic records
management systems”, “electronic archive management”, “digitalization of audiovisual archive heritage”, “private
archives and access”, “digital cultural heritage and access”, “electronic archives and research services”, “information
and records management training”, “user services in photo archives”, “corporate ERMS models”, “information and
records management” and “archival ethics and access”. When the results were evaluated, it was seen that important
topics in archival and records management were successfully discovered with the algorithm in question.