MACHINE LEARNING–DRIVEN PREDICTION OF COST OVERRUNS IN CIVIL CONSTRUCTION PROJECTS

Abstract

Cost overruns remain one of the most critical challenges affecting the successful delivery of civil construction projects worldwide. Inaccurate cost estimation, design changes, project delays, and unforeseen risks significantly contribute to budget escalation. This research presents a machine learning–driven framework for predicting cost overruns in civil engineering projects at early project stages. Historical project data comprising financial, technical, and managerial parameters are utilized to train predictive models. Multiple supervised learning algorithms are evaluated to identify the most reliable prediction technique. The proposed approach enables proactive decision-making and risk mitigation by forecasting potential cost overruns before project execution. Experimental results demonstrate improved prediction accuracy compared to conventional estimation methods. The study highlights the practical applicability of data-driven intelligence in construction cost management. The outcomes support enhanced financial control and sustainable project planning.