Using a three-stage evaluation procedure, this study
summarises a list of forecasting models that have been utilised
in the construction sector in the past. The research focuses on
publications published between 2012 and 2022 (all included),
with a total of 34 documents chosen for further examination.
The findings show that the number of publications varied
between 2012 and 2016, with a notable uptick after 2016. The
rising interest in this issue is shown by the distribution line from
2016 to 2020. The building cost index, shifting material prices,
and overall construction expenses were all major issues for the
researchers. The vector error correction model, the artificial
neural network, and the autoregressive integrated moving
average were the three most common forecasting models
used in the previous literature. Predictions relied heavily on
time series and ML methods. The hybrid model could be a
useful tool for researchers looking to enhance the precision of
their predictions in future investigations. The consequences of
natural catastrophes should also be considered in subsequent
research. Researchers and industry practitioners may use this
study's findings to choose an appropriate forecasting model
and put accurate prediction into practice.