Applying Artificial Intelligence and Human Influence Index for Predicting Cultural Heritage Risk in Wine-Growing Areas along the Danube River (Serbia)

Jovanović, Radmila and Manić, Emilija and Marin, Florentina (2026) Applying Artificial Intelligence and Human Influence Index for Predicting Cultural Heritage Risk in Wine-Growing Areas along the Danube River (Serbia). In: Proceedings of the 20th International Conference on Business Excellence (ICBE). Bucharest University of Economic Studies, Bucharest, Romania, pp. 2477-2485. ISBN 2558-9652

Text Jovanović, Manić, Marin - ICBE (2026).pdf - Published Version Available under License Creative Commons Attribution Non-commercial No Derivatives. Download (1MB)

Abstract

This study integrates artificial intelligence (AI) with the Human Influence Index (HII) to predict the risk of cultural heritage degradation. Six fortresses in the wine regions of the Danube basin in Serbia serve as case studies: Petrovaradin (Srem region), Belgrade/Kalemegdan (Belgrade region), Smederevo (Belgrade region), Ram (Mlava region), Golubac (Mlava region), and Fetislam (Negotin region). A hybrid methodology was developed combining geospatial HII assessment (population density, infrastructure, land use, nighttime lighting) with machine learning algorithms (Random Forest, Gradient Boosting Machine, Neural Networks) trained on multi-temporal data (1961–2024). Normalized HII values were used as input variables for the AI models. The Gradient Boosting Machine model resulted in 89% accuracy in risk classification. The results indicate that urban fortresses (Petrovaradin HII_norm=0.83, Belgrade HII_norm=0.87, Smederevo HII_norm=0.81) are increasingly threatened with High (2025) to Very High/Critical (2045) risk levels, while the protected fortress Golubac (HII_norm=0.24) remains at low risk. Critical tipping points were determined at which moderately threatened sites (Ram HII_norm=0.56, Fetislam HII_norm=0.52) become high-risk areas between 2033 and 2040. The AI-HII approach offers heritage managers predictive analytics to inform conservation investment priorities in wine tourism areas.

Actions (login required)

View Item