Identify Incident Factors to Support Aviation Safety Decision-Making: Proposing a Deep Learning Approach




Yang, Qiwei

Journal Title

Journal ISSN

Volume Title



Aviation is a complicated business, and safety is of paramount importance because casualties involve once an operational aircraft fails. Prevention is arguably the best strategy for aviation safety. Analyzing the database of past incidents to prevent accidents from happening has been approved a success. Incident reporting systems, a sub-system of business management systems, require human experts' efforts to review incident reports to identify probable primary and contributing factors of incidents so that new effective prevention plans can be designed accordingly. However, human experts' reviews have become prohibitively expensive nowadays because generated incident reports are increased exponentially due to the accelerated advancement in information technologies and the prosperity of the commercial and private aviation business. As a consequence, intelligent technologies shall be applied to help aviation experts facilitate incident analysis. The research presented in this study shows that the use of deep learning methods to automate incident report analysis is a promising approach. About 172,990 qualified incident reports from the ASRS database between January 1988 and July 2019 are collected to train and build a deep learning model to identify the primary and contributing factors for each incident report. The decision support solution I develop is not only automated and customizable, but has higher accuracy and better adaptability than conventionally machine learning methods in extant research. This novel application of the deep learning method on the incident reporting system can efficiently assist human experts in the incident analysis.


This item is available only to currently enrolled UTSA students, faculty or staff. To download, navigate to Log In in the top right-hand corner of this screen, then select Log in with my UTSA ID.


Deep Learning, Incident Report, NLP, Safety Management



Electrical and Computer Engineering