异常值检测方法在民航告警中的应用

随着民用飞机和航线数量的不断攀升,民航业进入了大数据时代,如何有效地检测异常数据并且避免它们对分析结果的不利影响是研究的焦点。本文围绕四分位法和目前民航界惯用的标准差法展开讨论,主要从数据集分布、异常值比例、异常值检测能力、抗扰动性和可视化效果5方面对四分位法和标准差法进行分析。最后通过快速存取记录器(Quick access recorder,QAR)数据分析论证,四分位法在上述5方面都表现出比标准差法更好的性质,四分位法更适合应用在民航告警领域。

Application of Outlier Detection Method in Civil Aviation Alarm

With the increasing number of civil aircraft and routes, the civil aviation industry has entered the era of big data, and how to effectively detect abnormal data and avoid their adverse effect on the analytical results is the focus of many researchers. This paper focuses on the quartile method and the standard deviation method which is popular in civil aviation, including the distribution of data set, proportion of outliers, outlier detection capabilities, anti-disturbance and visual effects. Finally, the quick access recorder (QAR) data analysis demonstrates that the quartile method has better properties and is more suitable for application in the field of civil aviation warning compared with the standard deviation method.