Multiclass Multiple-Instance Learning for Predicting Precursors to Aviation Safety Events
- 1 January 2022
- journal article
- research article
- Published by American Institute of Aeronautics and Astronautics (AIAA) in Journal of Aerospace Information Systems
- Vol. 19 (1), 22-36
- https://doi.org/10.2514/1.i010971
Abstract
In recent years, there has been a rapid growth in applying machine learning techniques that leverage aviation data collected from commercial airline operations to improve safety. Anomaly detection and predictive maintenance have been the main targets for machine learning applications. However, this paper focuses on the identification of precursors, which is a relatively newer application. Precursors are events correlated with adverse events that happen before the adverse event itself. Therefore, precursor mining provides many benefits, including the identification of factors relevant to the occurrence of an adverse event and their signatures, which can be tracked throughout a flight to alert the operators of the potential for an adverse event in the future. This work proposes using the multiple-instance learning framework, a weakly supervised learning task, combined with a carefully designed binary classifier leveraging a Multi-Head Convolutional Neural Network–Recurrent Neural Network (MHCNN-RNN) architecture. Multiclass classifiers are then created and compared, enabling the prediction of different adverse events for any given flight by combining binary classifiers, and by modifying the MHCNN-RNN to handle multiple outputs. Results obtained showed that the multiple binary classifiers perform better and are able to accurately forecast high-speed and high-path-angle events during the approach phase. Multiple binary classifiers are also capable of determining the aircraft parameters that are correlated to these events. The identified parameters can be considered precursors to the events and may be studied/tracked further to prevent these events in the future.Keywords
This publication has 21 references indexed in Scilit:
- Explaining Aviation Safety Incidents Using Deep Temporal Multiple Instance LearningPublished by Association for Computing Machinery (ACM) ,2018
- Multiple instance learning: A survey of problem characteristics and applicationsPattern Recognition, 2018
- Instability of functionally graded micro-beams via micro-structure-dependent beam theoryApplied Mathematics and Mechanics, 2018
- Using ADOPT Algorithm and Operational Data to Discover Precursors to Aviation Adverse EventsPublished by American Institute of Aeronautics and Astronautics (AIAA) ,2018
- Modeling Precursors for Event Forecasting via Nested Multi-Instance LearningPublished by Association for Computing Machinery (ACM) ,2016
- Anomaly detection via a Gaussian Mixture Model for flight operation and safety monitoringTransportation Research Part C: Emerging Technologies, 2016
- Challenges and Opportunities in Flight Data Mining: A Review of the State of the ArtPublished by American Institute of Aeronautics and Astronautics (AIAA) ,2016
- Detection of Precursors to Aviation Safety Incidents Due to Human FactorsPublished by Institute of Electrical and Electronics Engineers (IEEE) ,2013
- An overview of ensemble methods for binary classifiers in multi-class problems: Experimental study on one-vs-one and one-vs-all schemesPattern Recognition, 2011
- Error Prevention as Developed in AirlinesInternational Journal of Radiation Oncology*Biology*Physics, 2008