Prediction and Estimation of Highway Construction Cost using Machine Learning

Cost estimation and prediction are crucial processes for the success of construction projects, especially for infrastructure development. This study analyzes historical data collected between 2011 and 2023 and investigates the relationship between construction elements and the final cost of highway construction projects in Iraq. Different cost analysis approaches, including statistical assessment and machine learning techniques, were applied to a dataset of 291 highway projects. Cost estimation is a time-consuming and risky process that requires many qualitative and quantitative parameters to be well analyzed. However, machine learning provides a comprehensive assessment tool to predict future costs. Four ANN-based models were investigated and precision was improved by combining RMSE and the correlation coefficient (R) as a controller. The results showed improvements in performance metrics, such as error reduction rate and correlation coefficient, for the models developed. The best performance was achieved at an R of 0.989. The proposed model can be effectively adapted to predict road construction costs. Despite the need for more data, the implication of the proposed model can ensure a sustainable application, saving the time and resources required by construction professionals to predict road project costs during the planning phase.