Search

There are numerous models proposed for construction cost estimation. Most of them are based on projects' characteristics only while neglecting the external economic factors. This may be partially because there is no consensus on the effects of the economic factors on construction cost estimation and little attention has been paid to incorporating the trend of economic factors into cost estimation. More importantly, there is a general lack of quantitative analysis. To explore those effects quantitatively, this study uses deep neural networks (DNN) as an estimator and SHapley Additive exPlanations (SHAP) as a model interpreter, adopting the data on 98 public school projects in Hong Kong SAR. The analysis is also verified by a comparison analysis using several machine learning models popular in construction cost estimation. The results indicate that the economic factors do play an important role in reducing the construction cost estimation errors and are even more important than projects' characteristics. The findings would be helpful for stakeholders in the field of construction engineering and management to make appropriate decisions and for researchers to unveil the actual degree of the effects of other influential factors on construction cost estimation. • Economic factors can considerably reduce the errors of construction cost estimation. • Economic factors are more important than projects' characteristics in cost estimation. • The exploration is based on data on 98 public school projects in Hong Kong SAR. • DNN performs slightly better than the other machine learning tools, followed by RF. [ABSTRACT FROM AUTHOR]