Abstract:

In order to reduce railway violations and avoid railway traffic accidents, taking the observation records of unsafe behavior of Ganquan Railway in the past year as samples, data induction method and data analysis method were used to analyze the main causes of railway violations. The human factor model was used for analysis and comparison, and the action-type error model was selected as the analysis template. The correlation analysis tool decision tree C5.0 algorithm was introduced to establish the analysis model, the causal factors and correlation attributes were introduced, the neurons and contacts were distinguished, and the neural network graphic structure was constructed. After data analysis, it was concluded that the simplified operation process had the highest priority, and it was closely related to the employee's time in the enterprise and the employee's operation ability. The results show that: simplified operation process is the main cause of railway violations. Old employees with long working time and poor operation ability are more likely to simplify the operation process, and human errors mainly occur in the execution and planning stages. Railway enterprises should consider the analysis results of the model and develop targeted control measures to cut off the closed-loop path formed by railway violations, so as to prevent the occurrence of railway personal injury accidents.

Key words: railway operation, violation of regulations, human factor model, analysis of association, decision tree