Değerler anketi bağlamında dünya ülkelerindeki soyut gruplaşmaların değişiminin makine öğrenmesi teknikleri ile yıllara göre incelenmesi


Creative Commons License

Tezin Türü: Doktora

Tezin Yürütüldüğü Kurum: İstanbul Üniversitesi, Sosyal Bilimler Enstitüsü, Econometrics, Türkiye

Tezin Onay Tarihi: 2025

Tezin Dili: Türkçe

Öğrenci: SECA TOKER ASLAN

Danışman: Mehmet Hakan Satman

Açık Arşiv Koleksiyonu: AVESİS Açık Erişim Koleksiyonu

Özet:

The aim of this thesis is to produce a new alternative to the World Culture Map (WCM) developed by Inglehart and Welzel using Huntington's theoretical classification. In this thesis, the Integrated Values Survey (IVS) data, which is a combination of the World Values Survey (WVS) and the European Values Survey (EVS), was used. The IVS, which covers 115 countries worldwide between 1981 and 2021 and is a global sociological data set, consists of a total of 645,249 observations in 7 research periods. The size and comprehensiveness of the IVS data set made it possible to approach an alternative world culture map in a completely data-driven methodology. In this thesis, the Variable Selection Procedure (VSP), developed for each research period in the IVS dataset, was employed alongside a multi-factor structure, which reflects the dynamics of the entire period, unlike the two-factor structure used in WCM. Additionally, the unsupervised machine learning algorithm CLARA was utilized instead of Huntington's theoretical classification. In this context, a data-driven approach was adopted, as opposed to theoretical approaches, to analyze the cultural profiles of countries from a contemporary and dynamic perspective. In this context, a data-driven approach was adopted in contrast to theoretical approaches in order to analyze the cultural profiles of countries from a more dynamic and current perspective. This thesis aims to offer a more flexible and up-to-date perspective in a rapidly changing world with multicultural structures and cultural theories, in a process where cultural boundaries are being reshaped.