Train operation conflict detection for high-speed railways: a naive Bayes approach

Accurately detecting train operation conflicts (TOC) has great significance for improving the emergency handling ability of dispatchers during interference. In this study, a conflict detection model for high-speed train operation is proposed, with the train operation data from Xiamen to Shenzhen high-speed railway. Firstly, a TOC detection model framework considering data imbalance is determined, based on Bernoulli naive Bayes model. Then, the hyper-parameter of the proposed model is tuned with the training and validation dataset. Next, the performance result of the proposed model is compared to other three commonly used naive Bayes models, namely the Gaussian naive Bayes, multinomial naive Bayes and complement naive Bayes. Comparison analyses based on the commonly used classification model evaluation indexes show that the detection accuracy of the proposed model is significantly higher than other naive Bayes models. The proposed model also achieves high robustness and detection accuracy in each category.