Risk assessment of inland waterborne transportation using data mining

China has constructed a relatively complete inland waterborne transportation system. However, the frequent occurrence of inland water accidents with serious consequences, like the catastrophic Orient Star shipwreck, is an urgent unsolved problem. To reduce such accidents in the future and improve inland waterborne transportation safety, this study uses data mining, mainly containing text mining and association rule mining to risk assess China’s inland waterborne transportation, rather than the traditional quantitative risk assessment model. Text mining enables the risk factors to be objectively identified and distilled from accident reports. The potential relationships between risk variables are explored using association rule mining, based on the FP-Growth algorithm. The results reveal the essential problem facing China’s inland waterborne transportation system: frequent and varied ship accidents; key risk factors include overloading or improper loading, poor navigation visibility, inadequate sailor competence, and insufficient government supervision of shipowners and shipping companies. Combining the actual circumstances of inland waterborne transportation operations, this study proposes relevant recommendations for governments and relevant supervisory departments. The integrated application of text mining and association rule mining serves to avoid uncertainty and subjectivity, and achieve good results proving their scientific nature as a feasible method in water transportation risk research.