Aviation safety prediction method based on LSTM-RBF neural network model

Abstract

The invention discloses an aviation safety prediction method based on an LSTM-RBF neural network model. The aviation safety prediction method comprises the steps: S1, establishing the LSTM-RBF model;and S2, predicting the aviation safety by using the LSTM-RBF model; training cause event data samples by using an LSTM model, and analyzing the time sequence change trend of various cause events; an RBF model is used for training causal relationship between cause-causing events and aviation unsafe consequence event data, action weights of various cause-causing factors are calculated, and the modelpredicts the change trend of the aviation unsafe consequence events by researching the time sequence rule of the cause-causing events.

Description

technical field [0001] The invention relates to the technical field of computer processing, in particular to an aviation safety prediction method based on an LSTM-RBF neural network model. Background technique [0002] Safety is an eternal topic in the aviation industry. Aviation accidents will cause a large number of casualties, serious equipment damage, and serious social losses. Accurate aviation safety prediction is the prerequisite for effective safety early warning and prevention. At present, aviation safety prediction includes time series prediction, econometric model prediction and machine learning prediction, etc.; [0003] Time series forecasting is a linear forecasting model that uses time as an independent variable to conduct correlation analysis on samples to obtain the trend and cycle of data changes. Then identify a suitable random model and perform curve fitting. Commonly used time series forecasting models include AR, MA, ARMA and ARIMA models, etc. Since ...

AI Technical Summary

Problems solved by technology

However, the input of the traditional neural network model is generally considered to be independent of each other, that is, the time variable of the input layer is used as a general input variable, and the timing characteristics of the data sample itself are not considered.

This adds error to the forecast

Images