Risk influencing factors on the consequence of waterborne transportation accidents in China (2013–2023) based on data-driven machine learning

Abstract

Early warning on the basis of RIFs data is widely considered as an effective way to prevent waterborne transportation accidents, and the performance of warning model is critical. To develop a warning model with good performance, in this study, a data-driven based comprehensive machine learning algorithm, namely BiLSTM-CNN-RF is proposed. The RIFs data used to train the proposed algorithm is extracted from the 1090 waterborne transportation accident investigation reports during 2013–2023 in China, the collected data is first pre-processed to establish the input sample set of the algorithms. Meanwhile the importance of RIFs is also quantitatively analyzed. The traditional machine learning algorithms, such as RF, SVM, MPL, and GRU, are also involved in this study to verify the performance of the proposed comprehensive algorithm. The RIFs data is then fed into these five machine learning algorithms, the prediction results of “Accident type” and “Accident grade” are used to examine their prediction performance. The results show that the performance of the proposed BiLSTM-CNN-RF algorithm is better than the four traditional machine learning algorithms, especially for prediction accuracy, and another superiority is the good applicability in case of small sample data volume.