Time series analysis of construction Cost Index using wavelet transformation and a neural network

Abstract

Construction Cost Index (CCI) is widely used to analyze the construction cost variation in time. It can convert the present construction cost to the future or to the past. The current practice of the future cost estimation is simply an extrapolation of the past CCI, which often leads to the inaccurate estimation of future construction cost. This paper presents a new CCI forecasting model using wavelet transformation and an artificial neural network. Preliminary tests showed that the proposed method produced the shortterm range of a future CCI with a greater accuracy as well as a higher reliability compared to existing methods.