Estimating Construction Project Duration Using a Machine Learning Algorithm

Abstract

Construction project delays remain one of the most relevant problems in the construction sector. The construction industry is also one of the least digitalized industries. This research aims to use the power of artificial intelligence and machine learning to help better understand the delays faced in building construction projects and be able to estimate them before the project begins. A thorough literature review was conducted to find the main causes of construction delays in Lebanon. A model was then created using the machine learning algorithm extreme gradient boosting (XGBoost) based on factors that quantify the main causes of delay that can be known before the project begins. The goal of the model is to be trained on the training data to accurately predict the delay of projects that were not seen by the model before. Previous research into construction project delays has only created models that classify projects by their delay risk level. No research has been done on the use of machine learning to create regression models that can predict the delay of a project before the project starts. This research fills the gap by creating a model that can estimate construction project delays before projects begin. The model estimated project delays with an error of 24% and an adjusted R² of 74.3%. This shows that the model was able to achieve relatively accurate results and explain 74.3% of the variability of the delay while only using ten factors causing delay. The results show that the factors mostly affecting delay in Lebanese construction projects are the client’s performance, legal issues faced by the project, the project manager’s expertise, and the quality of design documents.