Estimation of Building Construction Cost Using Artificial Neural Networks

Abstract

The cost estimation of the building construction projects at initial stages with a higher degree of accuracy plays a vital role in the success of every construction project. Based on the survey and feedback of the design professionals and construction contractors, a dataset of 78 building construction projects was obtained from a mega urban city Mumbai (India) and geographically nearby region. The most influential design parameters of the structural cost of buildings (Indian National Rupees: INR) were identified and assigned as an input and the total structural skeleton cost (INR) signifies the output of the neural network models. This research paper aims to develop a multilayer feed forward neural network model trained along with a backpropagation algorithm for the prediction of building construction cost (INR). The early stopping and Bayesian regularization approaches are implemented for the better generalization competency of neural networks as well as to avoid the overfitting. It has been observed during the construction cost prediction that the Bayesian regularization approach performance level is better than early stopping. The results obtained from the trained neural network model shows that it was able to predict the cost of building construction projects at the early stage of the construction. This study contributes to construction management and provides the idea about the entire financial budget that will be helpful for the property owners and financial investors in decision making and also to manage their investment in the volatile construction industry.