An effective method for identifying the key factors of railway accidents based on the network model

Abstract

In the analysis of railway accidents, the identification of key factors is an essential step for effective monitoring and daily maintenance of railway system. The existing key factor identification methods are generally based on the importance of ranking, which cannot achieve effective identification and verification because the classification of accident causal factors is not considered. Based on the complex network model, this paper proposes a heterogeneous network model which is suitable for the classification of accident causal factors to find the key causal factors of railway accident. The proposed model considers the causal factors of different types of railway accidents separately. Here, the accident causal factors are taken as the nodes, and their relationships are considered as the edges. The key factors will be determined through the extraction of the key nodes. Furthermore, an improved SIS propagation model is used to verify the effectiveness of the proposed method. Because of the strong network heterogeneity, it is proved that the proposed method is suitable for finding the key nodes in heterogeneous networks. Morever, the results show that the key nodes found by using the proposed method have more important impacts due to their faster propagation in the SIS model.