Risk Identification and Prediction for Highway Bridge Projects Using an Artificial Intelligence Model

The risk management is quite important for the highway bridge projects due to the complicate structures and technologies influenced by site conditions, environment, service condition, etc. The risks in these projects are sometime unavoidable, so the identification and prediction of risk are required to ensure the high requirements of safety, quality, longevity and to optimize the management and maintenance activities. Technical risks of bridge in Vietnam have identified and analyzed by some authors with initially proposing of effective models and solutions to control and limit the risks. However, the risk factors in bridge projects are comparatively fuzzy, random and complicatedly connected to each other. Hence, there are some disadvantages of these proposed models and methods in solving the complex and multi-risk problems and the prediction has not been well done. In this paper, an Artificial Intelligence model for technical risk identification and prediction in highway bridge projects will be constructed based on the previous investigation results during the bridge construction and operation in Vietnam, the ensemble methods with Random Forest algorithm will be discussed and be used to obtain better risk predictive performance, and then, the application for real projects is intended.