Cost Prediction for Roads Construction using Machine Learning Models


  • Yasamin Ghadbhan Abed University of Diyala, College of Science, Department of Computer Science, Diyala, Iraq
  • Taha M. Hasan University of Diyala, College of Science, Department of Computer Science Diyala, Iraq
  • Raquim Nihad Zehawi University of Diyala, College of Engineering, Department of Highway and Airport Engineering Diyala, Iraq



Construction, Roads, Cost estimation, Machine learning, Ridge regression


Predicting conceptual costs is among the essential criteria in project decision-making at the early stages of civil engineering disciplines. The cost estimation model availability that may help in the early stages of a project could be incredibly advantageous in respect of cost alternatives and more extraordinary cost-effective solutions periodically. There is a lack of case datasets. Most of the proposed dataset was inefficient. This study offers a new data set that includes the elements of road construction and economic advantages in the year of project construction. Real project data for rural roads in the State of Iraq / Diyala Governorate for the years 2012 to 2021 have use to train a predictive model with a high rate of accuracy based on machine learning (ML) methods. Ridge and Least Absolute Shrinkage and Selection Operator (LASSO) Regressions, K Nearest Neighbors (k-NN), and Random Forest (RF) algorithms have employ to create models for estimating road construction costs based on real-world data. The Root Mean Square Error (RMSE), Mean Absolute Percentage Error (MAPE), and R-squared (R2) coefficient of determination are utilize to assess the models' performance. The analysis indicated that the RR is the best model for road construction costs, with results R2 = 1.0, MAPE =0.00, and RMSE=0.00. The results showed that the cost estimates were accurate and aligned with the project bids.




How to Cite

Y. G. Abed, T. M. Hasan, and R. N. Zehawi, “Cost Prediction for Roads Construction using Machine Learning Models”, IJECES, vol. 13, no. 10, pp. 927-936, Dec. 2022.



Original Scientific Papers