Machine Learning Application for Automatic Analysis of Aircraft Operation Interruption Reports

摘要

1.ML/NLP application for field report automatic classification using supervised machine learning algorithms. 2.Responsiveness as main benefit. 3. Results presented satisfactory accuracies ranging from 78% to 90%. 3.Performance was compromised by inconclusive reports samples at the historical database as well as poorly populated labels. 4.Solution highly dependent on reliable historical data (application for new aircraft models may be an issue).