Bayesian network model for stakeholder management in large scale housing projects: prediction of project success

Abstract

Stakeholder management in construction projects affects cost, time, quality, and safety which are considered as the success parameters of projects. The aim of this study is to develop a predictive model using a Bayesian Network (BN) approach to measure the likelihood of project success, based on the general contractor's stakeholder management framework. This study consists of two stages; in the first stage Key Stakeholder Mapping was created using expert panel. In the second stage, BN Model was developed with 64 large-scale housing project data. BN is formulated as a probabilistic predictive model aiming to assess the impact of key stakeholders on projects. This model enables the main contractor to identify key stakeholders' position over the project and generate an objective base to develop strategies to decrease their adverse effect over the success of the project. The accuracy rate of the overall model is 63%, prediction on compliance with the project cost is 84% and prediction on safety/accident rate is 80%. The study states that a comprehensive review of the likelihood of success will help in identifying and addressing crucial uncertainties within the project; this will contribute to the development of scenarios to improve the project's probability of success.