Skip to content
Conference Open access

A Machine Learning Framework for Predictive Modeling of Construction Project Duration Using Data-Driven Intelligent Methods

Sep 2026 · 2026 11th International Conference on Computer Science and Engineering (UBMK) · pp. 935-940 · 0 citations · 7 references

Abstract

Construction project duration is considered a critical challenge due to the complex interactions between technical, environmental, and management variables that exist in all projects. The present study develops an intelligent approach to predict the duration of construction projects based on machine learning methods. A dataset comprising 300 real construction projects implemented in Jordan was utilized in this study. The proposed technique includes the steps of data preprocessing, evaluation of baseline and artificial neural networks (ANN) models, train/test evaluation, five-fold cross-validation, hyperparameter tuning, and feature importance assessment. At first, eight regression algorithms were tested to find out the most suitable one for this problem. Then, an ANN model was applied as a meta-learner to combine predictions generated by multiple algorithms into features and use them as input for the ANN training process. Model performance was assessed using the coefficient of determination (R2), mean squared error (MSE), and mean absolute error (MAE). The results show that Linear Regression achieved the highest and most stable performance, with very close training and testing R2 values of 0.988658 and 0.986788, respectively, indicating excellent generalization. In comparison, the best ANN model, ANN (20, 20), achieved a lower testing R2 of 0.922874, showing competitive but less accurate performance than Linear Regression. Feature-importance and explainability techniques were applied to interpret the contributions of the input variables to project-duration prediction. The results showed that Linear Regression achieved the highest accuracy over the tested models.

Read PDF

We use cookies to run the site and, with your consent, for analytics and to show ads. See our Cookie Policy.