Open AccessNeural network-based probability forecasting method of aviation safety
Author(s) -
Zheqi Zhu,
Bo Ren,
Xiaofeng Zhang,
Hui-yong Zeng,
Tao Xue,
Chen Qingge
Publication year - 2021
Publication title -
iop conference series. materials science and engineering
Language(s) - English
Resource type - Journals
eISSN - 1757-899X
pISSN - 1757-8981
DOI - 10.1088/1757-899x/1043/3/032063
Subject(s) - aviation , probabilistic forecasting , aviation safety , civil aviation , artificial neural network , probability distribution , interval (graph theory) , range (aeronautics) , aviation accident , computer science , operations research , engineering , econometrics , statistics , artificial intelligence , mathematics , probabilistic logic , combinatorics , aerospace engineering
Aviation safety forecasting is of great significance for accident prevention. At present, aviation safety forecasting is mainly deterministic forecasting, ignoring the impact of various uncertainties on forecasting. In this paper, on the basis of deterministic forecasting, the forecasting of aviation safety probability is carried out based on the uncertainty of neural network point forecasting value. The uncertainty of aviation safety forecasting is described by three ideas: the numerical statistical characteristics of point forecasting value, the probability density fitting of point forecasting value and the distribution of error. The interval forecasting result is obtained, which can better understand the uncertainty and risk of forecasting quantity. Taking the aviation safety data of civil aviation from 1993 to 2008 as an example, the results show that this method can provide the most possible range of aviation safety forecasting under certain uncertainty level, and is more conducive to the modeling of aviation safety uncertainty.