Akademik Veri Yönetim Sistemi
Mining, Metallurgy and Exploration, cilt.43, sa.2, ss.1007-1020, 2026 (SCI-Expanded, Scopus)
Türkiye’s recent increasing popularity in the global natural stone market necessitates accurate and reliable predictive modelling of its export performance to support national economic stability and sectoral planning. In addition, the highly dynamic nature of global trade, coupled with fluctuations in domestic economic factors, introduces significant complexity and non-linearity, thereby challenging traditional linear prediction approaches. Accordingly, this study addresses this gap by applying a power-law-based multiple nonlinear regression (MNLR) framework to empirically model Türkiye’s natural stone exports (NSE). The developed MNLR model integrates five critical macroeconomic indicators as independent variables, namely gross domestic product (GDP), unemployment rate (UP), interest rate (IR), exchange rate (ER) and inflation (INF). To evaluate the reliability and predictive performance of the model, well – known error indicators—including root mean square error (RMSE), relative root mean square error (RRMSE), correlation and determination coefficients (r and R2), and variance inflation factor (VIF)—are employed. Moreover, the most influential indicators affecting NSE performance are identified. The results demonstrate that the proposed model, based on the MNLR approach, can serve as an effective analytical tool for assessing nonlinear relationships, supporting strategic decision-making, informing policy and investment strategies, and enhancing the competitiveness of Türkiye’s natural stone sector in global markets.