Integrating Professional Judgement with Machine Learning Models for Enhanced Construction Cost Estimation and Prediction in Contextually Variable Projects

About the Project

Construction cost estimating and prediction have been modelled using stochastic and parametric approaches. Algorithmic tools have been developed to improve precision and accuracy of estimates and forecasts of construction costs. Yet construction cost prediction remains dynamic in the sense that each project is unique not just in design and construction, but also in the prevailing client and environmental context. Consequently, it is understood that professional judgement sits at the centre of the subjective processes of selecting cost/market data inputs, estimating and forecasting processes used, and evaluation of outputs. Attempts have been made to model professional judgement in this context in the past. However, the potential for more accurate cost predictions that bring together professional evaluation of the variable contexts specific to the project and the technologically developed estimates and forecasts based on historical cost data may need to be explored.

The research is expected to address the gap between algorithmic cost estimation methods and the nuanced, context-sensitive judgement of professionals and to propose a hybrid approach that leverages machine learning for data-driven predictions while incorporating expert insights to adapt to the unique characteristics of each construction project.

Funding Notes

there is no funding for this project