基于改进贝叶斯神经网络的航空公司安全敏感性分析

针对当下航空公司安全质量管理体系(Quality Management System，SQMS)中风险识别与可靠性改进的问题，提出了基于区间数学改进的贝叶斯神经网络的灵敏度分析方法。利用区间数学理论分析贝叶斯神经网络中各指标与整体安全质量状况的扰动关系，实现指标灵敏度分析。通过东方航空公司的实例分析，发现在对指标进行人工干预时组合指标干预效果较好，且安全管理体系实施后指标的灵敏度有明显向好的方向变化的趋势。

Analysis on safety sensitivity of airline companies based on improved Bayesian neural network

To meet the requirements of risk identification and reliability improvement on Safety and Quality Management System (SQMS) in airline companies, a method of sensitivity analysis by improved Bayesian neural network based on interval mathematics was put forward. The interval mathematics theory was applied to analyze the disturbance relationship between each index and overall safety and quality status of Bayesian neural network, so as to realize the sensitivity analysis of indexes. The case study of Eastern Airlines showed that the intervention effect of combined indexes was better when conducting artificial intervention to indexes, and the sensitivity of indexes tended to be better obviously after the implementation of safety management system.