Scholarlink

<p>The aim of this study is to offer a comparative world examination of countries by economic freedom metrics and to evaluate structural contrasts using machine learning methodology. Economic freedom is a complex characteristic with institutional and market dimensions such as property rights, judicial independence, government honesty, and business freedom. Principal Component Analysis (PCA) and K-Means clustering were employed to identify concealed structures from the data set and discovered noteworthy groupings between countries. These groups were again validated by supervised learning algorithms, including Random Forest, Support Vector Machines, Decision Tree, XGBoost, and Artificial Neural Networks. Amongst these, Random Forest classifier yielded the highest classification accuracy (97.3%) and AUC score (0.992), followed by XGBoost (96.4%, AUC: 0.987) and ANN (95.1%, AUC: 0.981). The findings demonstrate that business freedom is the greatest economic freedom determinant, followed by governance quality, property rights, and government integrity. In contrast to classical fiscal indicators, tax burden and government spending proved to be smaller determinants. The models’ good classification performance indicates that the economic freedom groupings possess good structural consistency and learnability. Results of the study help policymakers with practical lessons by highlighting priorities of reforms in economic freedom and guiding the planning of sustainable development policy.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Multidimensional analysis of global economic freedom performances: structural findings and machine learning approaches

  • 期刊论文

  • 发表日期：

    2026年1月28日

查看全文
  • Sadullah Çelik

摘要

The aim of this study is to offer a comparative world examination of countries by economic freedom metrics and to evaluate structural contrasts using machine learning methodology. Economic freedom is a complex characteristic with institutional and market dimensions such as property rights, judicial independence, government honesty, and business freedom. Principal Component Analysis (PCA) and K-Means clustering were employed to identify concealed structures from the data set and discovered noteworthy groupings between countries. These groups were again validated by supervised learning algorithms, including Random Forest, Support Vector Machines, Decision Tree, XGBoost, and Artificial Neural Networks. Amongst these, Random Forest classifier yielded the highest classification accuracy (97.3%) and AUC score (0.992), followed by XGBoost (96.4%, AUC: 0.987) and ANN (95.1%, AUC: 0.981). The findings demonstrate that business freedom is the greatest economic freedom determinant, followed by governance quality, property rights, and government integrity. In contrast to classical fiscal indicators, tax burden and government spending proved to be smaller determinants. The models’ good classification performance indicates that the economic freedom groupings possess good structural consistency and learnability. Results of the study help policymakers with practical lessons by highlighting priorities of reforms in economic freedom and guiding the planning of sustainable development policy.