•  
  •  
 

Faculty Mentor

Burak Cankaya

Abstract

Aircraft incident data from the Federal Aviation Administration between 1978-2023 was analyzed for the top five airlines in the United States: Delta Air Lines Inc., American Airlines Inc., Alaska Airlines Inc., Southwest Airlines Co., and United Airlines Inc. This analysis aimed to predict conditions with a higher chance of causing injury and determine the best machine-learning model for this prediction. The target variable of "injuries occurred" is processed through machine learning methods such as Naïve Bayes, logistic regression, and XGBoost. This was to predict the injuries and understand and explain the complex fault mechanism with explainable artificial intelligence. The XGBoost model gave the best predictive performance at 93% receiver operating characteristic, 84.4% precision, and 88.2% recall. This provided insights into improving safety in the aviation industry. It contributed to the literature by predicting injuries in legacy airlines, explaining the patterns that create incidents, and creating prescriptions to reduce injuries that happen on airlines using explainable machine learning and text-mining models.

Share

COinS
 
 

To view the content in your browser, please download Adobe Reader or, alternately,
you may Download the file to your hard drive.

NOTE: The latest versions of Adobe Reader do not support viewing PDF files within Firefox on Mac OS and if you are using a modern (Intel) Mac, there is no official plugin for viewing PDF files within the browser window.