基于PCA-GA-BP的民航风险评价体系和预测模型

Civil Aviation Risk Evaluation System and Forecast Model Based on PCA-GA-BP

近年来我国民用航空处于稳步发展阶段,发展中的安全问题需要得到高度重视。目前对于民用航空安全的预测均需大量数据,且收集数据较为困难。针对这一问题采用了PCA-GA-BP组合模型,对指标之间的相关性进行分析,从而优化了风险评价体系,对数据进行了降维,减少了所需数据数量;以收集到的从2012年1月至2015年4月共40个月的其他不安全事件数据为输入,以事故征候率这一较为稳定且与安全水平联系紧密的指标为输出对象,建立了民用安全预测模型。算例表明,与PCA-BP模型91.885%的准确率相比,PCA-GA-BP模型的准确率达到93.444%且有更好的稳定性;与GA-BP模型相比,PCA-GA-BP模型在不降低预测精确度的前提下将运行速度提高近50%,为民用航空安全预测提供了新思路。

In recent years,China's civil aviation has been in a steady development stage,and safety issues in the development need to be given high attention.At present,the prediction of civil aviation safety requires a large amount of data,and collecting data is difficult.This paper uses the PCA-GA-BP combination model to analyze the correlation between indicators,thereby optimizing the risk assessment system,reducing the dimensionality of the data,reducing the amount of data required,and establishing an acci⁃dent rate prediction model.This article takes the data collected from other unsafe events for a total of 40 months from January 2012 to April 2015 as input,and uses the incident rate,a relatively stable indicator that is closely related to the safety level,as the out⁃put to establish the civil security prediction model.The calculation example shows that compared with the 91.885%accuracy of the PCA-BP model,the accuracy of the PCA-GA-BP model reaches 93.444%and has better stability.Compared with the GA-BP mod⁃el,the PCA-GA-BP model can increase the operating speed by nearly 50%without reducing the accuracy of the prediction,provid⁃ing new ideas for civil aviation safety prediction.