Generic Deep-Learning-Based Time Series Models for Aviation Accident Analysis and Forecasting

The global civil aviation is growing day by which is resulting into more congestion into the air transportation system. Over the past few years, many efforts have been put in the direction to improve the aviation safety and avoid fatal air accidents. Learning from the previous accidents could be crucial to enhance the safety standards and minimise the possibility of such aviation safety events. Hence, this manuscript provides an analysis of time-series-based machine learning models to investigate the impact of aviation accident injuries and fatalities. Time series models are highly capable of predicting aviation accidents due to their statistical association. This study has examined ASRS database containing accident and incident data from 1988 to 2020 by employing time series models such as autoregressive integrated moving average (ARIMA), seasonal ARIMA with an exogenous factor (SARIMAX), deep-learning-based long short-term memory (LSTM), bidirectional LSTM (Bi-LSTM), and gated recurrent unit (GRU). Here, MSE, RMSE, and MAE are used as performance indicators. The results proved that the Bi-LSTM prediction was superior compared to the other models. Deep-learning-based analysis helps to bring forward the various hidden patterns along with essential information for accident prevention and reducing the accident probability for improving the aviation safety. Ultimately, overall aviation safety level will enhance by minimising the likelihood of occurrence of unsafe events and aviation management optimise their actions towards safety © 2023, The Author(s), under exclusive licence to Springer Nature Singapore Pte Ltd.