Optimizing Project Cost and Duration Forecasting through Machine Learning

The ability to accurately predict the duration and final costs of a project is essential to implement timely corrective actions during its execution. Estimates at Completion (EAC) are the key tool to support project control. In recent years, the integration of Machine Learning (ML) techniques with traditional methodologies such as Earned Value Management (EVM) and Earned Schedule (ES) has improved the accuracy of forecasting. However, many studies have focused exclusively on static data, neglecting both the influence of dynamic data and topological indicators and the issues of underfitting and overfitting that compromise model robustness. This aims to address both issues by proposing a structured ML pipeline, including automated preprocessing, feature engineering, and cross-validation procedures, with the goal of improving the generalizability of predictive models. The pipeline is designed to exploit both static data and dynamic data and topological network indicators by analyzing their impact on predictive performance. The pipeline was tested using 30 machine learning algorithms on a dataset of 90 real projects by evaluating their effectiveness with respect to mean square error, project progress stage, and via SHAP analysis for interpretability. The findings indicate that ML algorithms outperform classical approaches concerning accuracy and precision particularly during the initial and mid phases of a project. Moreover, feature analysis with SHAP underscored the tremendous value of dynamic data and project network attributes concerning model prediction capability. In conclusion the study demonstrates the effectiveness of the proposed pipeline and suggest that further integration of ML in project management practices could lead to improved project outcomes, especially as ML techniques continue to evolve.