Study on Test Method for Outlier of Flight Data

This paper applies RBF (Radial Basis Function) neural network to the outlier test of flight data based on the features of flight data. It tests the outlier situations of the output target parameters by establishing the reasonable functional relationship model as well as fitting the functional relationship between the input parameter and the output parameter of the aircraft engine system. Considering the interaction among parameters in various systems of the aircraft, this method tests the outlier situations of the target attribute by fitting certain functional relationship from correlated multidimensional attributes. It gains relatively good effects by experimental analysis, providing support for the study on fault diagnosis and trend prediction by utilizing flight data.