Title: A predictive analytic approach to determine construction cost estimates

Authors: Asil Oztekin; Rory R. Masterson

Addresses: Manning School of Business, University of Massachusetts Lowell, One University Avenue, Southwick 201 Operations and Information Systems, Lowell, MA 01854, USA ' Manning School of Business, University of Massachusetts Lowell, One University Avenue, Southwick 201 Operations and Information Systems, Lowell, MA 01854, USA

Abstract: In the construction industry, cost estimates are the basis for which projects are awarded to suppliers. A cost estimate is arguably the most important inclusion in a supplier's response to a customer's 'request for quote'. As such, it is essential for construction cost estimates to include features of accuracy, cost effectiveness and profitability for the supplier. The amount of variables and qualitative considerations in an estimate require a methodology beyond an objective approach. This paper uses data analytics to analyse both a US Government service provider's internal estimate process and suppliers' external estimate process to determine the best approach to use on future task orders. Three different classifier methods were used to perform the data analytic approach. Based on analyses performed for all three methods, RBF-based support vector machine with a 66%-33% random split and ten-fold cross-validation yielded the highest accuracy at 71%.

Keywords: construction estimates; decision support systems; artificial intelligence; decision tree; neural networks; support vector machines; SVMs; business analytics.

DOI: 10.1504/IJOR.2018.092013

International Journal of Operational Research, 2018 Vol.32 No.2, pp.169 - 200

Received: 18 Mar 2015
Accepted: 18 May 2015
Published online: 30 May 2018 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article