Search

AbstractThis paper summarizes a list of previously used forecasting models in the construction industry, using a three-stage review process. Specifically, articles published between 2012 to 2022 (inclusive) are investigated, and 34 documents are selected for further analysis. The results present a fluctuating number of publications from 2012 to 2016 and a significant increase since 2016. The distribution line between 2016 to 2020 indicates the growing interests in this topic. In addition, fluctuating material prices, construction costs and construction cost index were top concerns of researchers. Previous literatures frequently employed three forecasting models: Vector error correction model, Artificial Neural Network and Autoregressive Integrated Moving Average. Time series and machine learning techniques were widely used in prediction. Future studies could consider the hybrid model as a research method to improve the forecast accuracy. Additionally, future studies should take the effects of natural disasters into account. Findings of this study are expected to help researchers and industry practitioners select a suitable forecasting model and implement precise prediction in practice.