APPLYING MACHINE LEARNING FOR PREDICTIVE ANALYSIS IN PROJECT-BASED DATA: INSIGHTS INTO VARIATION ORDERS

The complexity of the global supply chain and project execution necessitates advanced methodologies in project management. As industries are generating large amounts of project data, machine learning (ML) algorithms can be a viable tool for addressing predictive analytics and transforming this industry into more digitalization. This study examines the feasibility of leveraging ML models for predicting variation orders (VOs) in an energy construction project through the use of actual project management data. Using historical project data, this study presents the investigative analysis of applying six ML regression models to predict VOs and evaluates the performance of these models using the mean squared error metric. It is observed that various project activities are nonlinear in the impact of the order of variation, which indicates that advanced ML techniques are required when analyzing the order of variation rather than using linear model analysis. Thus, the results underscore the critical role ofML predictive model implementation in improving change management by enabling preemptive detection of potential problems, risk reduction, and more efficient project execution. Moreover, this study will also help to narrow the existing gap between ML-based theoretical applications and practical project management strategies while also demonstrating the efficacy of AI-based decision support systems for on-time project control. The contributions of this study provide a foundation for developing integrated ML models and project management software, fostering data-driven decision making in dynamic project scenarios.