Institute of Transportation, MOTC

In recent years, there are few domestic and foreign research literatures relating to the delayed entry ships. To solve the problem of delayed entry containerships in the situation of multiple carriers, multiple ports and multiple terminals, the Bayesian network model is used in this study. The research data is in-port ship lists of Keelung Port, Taipei Port, Taichung Port, and Kaohsiung Port. Two types of models are built through the Bayesian network. Finally, the cross-validation is used to evaluate the accuracy and predictive ability of the model. The results show that the Bayesian network model can effectively predict the possibility of the entry delays and the delay level. The entry delays of the container ships are mainly related to the “carriers” of the container ships and the entry “week”. According to this study, port units can re-examine the entry procedures of container ships. In addition, achieving the optimal time planning and space utilization in the port.