Predicting conceptual cost for field canal improvement projects

A conceptual cost estimation is prepared to assess the feasibility of a project or establish the project's initial budget at the early stages of the project. The main objective of the paper is automating the cost estimate at the conceptual stage with the highest accuracy. The key contribution of this paper is developing a quadratic regression model with a prediction accuracy of 9.12% and 7.82% for training and validation, respectively. This research has identified the model's key parameters to establish a reliable conceptual cost estimate model for field canal improvement projects (FCIPs). Two machine learning models were developed utilizing multiple regression analysis (MRA) and artificial neural networks (ANNs). Searching for a better model, several data transformations have been conducted to improve the model performance. The quadratic regression model has shown the highest performance based on the correlation and the mean absolute percentage error (MAPE) criteria. A parametric model has been presented in this paper to predict the conceptual cost of FCIPs. This research maintains the importance of identifying key parameters and conducting data transformation and sensitivity analysis for developing a reliable parametric cost prediction model.