Conceptual Cost Estimation In Green Buildings By Using Regression Analysis And Artificial Neural Network Methods To Improve Accuracy

Abstract

Conceptual cost estimation is a critical task during the early stages of a construction project, especially for green buildings, which present unique sustainability challenges and design complexities. Traditional methods such as regression analysis are widely used but often rely on experienced estimators and are time-consuming, while Artificial Neural Networks (ANN) offer a modern alternative but are limited by the quality and quantity of available data. This study aims to develop a hybrid model combining regression analysis and ANN to improve the accuracy of conceptual cost estimation for green and conventional high-rise buildings in Indonesia. Using data from 22 high-rise building projects (13 conventional and 9 green buildings), the study employed regression analysis, ANN techniques, and a combination of the two, with eight key variables selected for modeling. The hybrid model demonstrated the highest accuracy, achieving a Mean Absolute Percentage Error (MAPE) of 15.09%, within the acceptable range for conceptual cost estimation (+10–30%) as per AACE standards, outperforming standalone regression and ANN models. These findings highlight that integrating regression analysis and ANN provides a robust tool for early-stage cost estimation, supporting sustainable construction practices and informed decision-making.