A new approach to improving the responsiveness to price fluctuations of the range estimating model through autocorrelation time-series analysis

Dynamic and volatile characteristics are the most significant factors that distinguish the construction industry from other fields. Many researchers have endeavored to deal with uncertainty caused by the above factors, and significant outcomes have been obtained, owing to the innovative methodologies of data handling. With the help of powerful data processing, enormous information, which is effective in coping with uncertainty, could be gained. During the last three decades, the housing industry in Korea has boomed, resulting in huge data generation. More precise estimation is required at the initiative phase to support decision-making on the possibilities of realization of projects. For more accurate estimation, a Range Estimating Model based on time series analysis is developed and suggested. This paper is organized into three main sections. In the first section, unit prices of residential building projects turn out to be time-dependent, from analysis of the Durbin-Watson ratio. The second section explores i) analysis of stationarity, ii) model development, and iii) model validation and application. In this process, this paper suggests appropriate time series models, such as the ARIMA, and Monte Carlo Simulation using the predicted unit prices. In order to validate the proposed model, priced bills of quantities of 150 housing projects are analyzed, and the results of a t-test on relative accuracies indicates that the proposed model is more accurate than the conventional range estimating technique, using historical cost data, and ignoring price fluctuations.