Dynamic-Bayesian-Network-Based Project Cost Overrun Prediction Model

One common problem in the construction industry is project cost overrun. Cost overrun can have significant impacts on financial profitability, completion, quality, and stakeholder satisfaction. The average percentage of vary widely depending type, size, complexity, location. Many approaches be adopted to prevent or mitigate overrun, one them a more accurate estimate prediction. Several studies estimation prediction been conducted based historical data; nevertheless, each has its characteristics trend. Real-time, project-specific data are reliable for forecasting trend itself. There many influence factors that may interdependently affect This paper proposes real-time predict risk model their interdependence as well corrective actions if adopted. study used dynamic Bayesian network (DBN) formulate architecture use input–output hidden Markov method (I/O HMM) with particle filter (PF) run inference. Six building mass rapid transit projects Taiwan were validation comparison. posterior probabilities from DBN-based highly consistent ratios real projects. Moreover, it superior other models terms accuracy. proposed could provide managers an early alert