<![CDATA[Unit price escalation in Water Resources (SDA) projects is a challenge for project implementation success. The uncertainty of material prices, labor wages, equipment costs, energy prices, inflation, exchange rates, changes in government policies, spatial conditions, project characteristics, and various external risks make price escalation predictions increasingly complex and difficult to perform using conventional approaches. The development of ML combined with DSS offers a more adaptive, accurate, and data-driven predictive approach to support decision-making of projects. This study aims to examine the application of ML in predicting unit price escalation in SDA projects, identify factors that influence prediction accuracy, and formulate directions for integrated DSS development. The study uses the Systematic Literature Review (SLR) method based on the PRISMA 2020 guidelines by reviewing articles published in the period 2015–2025 from the Scopus, Web of Science, ScienceDirect, IEEE Xplore, and SpringerLink databases. The results of the study indicate that the Random Forest, Extreme Gradient Boosting (XGBoost), Artificial Neural Network (ANN), Long Short-Term Memory (LSTM), and Gradient Boosting algorithms provide better predictive performance than conventional statistical methods. Most studies are still oriented towards general project cost estimation and have not developed a DSS capable of dynamically integrating economic factors, labor, project characteristics, spatial conditions, risks, and historical data to predict unit price escalation of natural resource projects. ]]>
Copyrights © 2026
