Phase of Flight Classification in Aviation Safety Using LSTM, GRU, and BiLSTM: A Case Study with ASN Dataset

2023-12-06

2023 International Conference on High Performance Big Data and Intelligent Systems (HDIS)

AI Summary

Identifying flight phases from unstructured aviation accident narratives remains challenging due to implicit temporal and operational cues. Method: This study proposes a deep learning–based automated classification framework using LSTM, GRU, and BiLSTM architectures—along with their pairwise hybrid variants (GRU-LSTM, LSTM-BiLSTM, GRU-BiLSTM)—integrated with text preprocessing and sequence modeling techniques. Contribution/Results: It presents the first systematic comparison of standalone versus hybrid RNN models for this task: BiLSTM achieves 64% accuracy, while the optimal hybrid model (LSTM-BiLSTM) reaches 67%, significantly outperforming baseline methods. Results demonstrate that deep learning can effectively infer latent flight phases from accident reports; moreover, synergistic integration of bidirectional context and gating mechanisms enhances representational capacity. This work establishes a reproducible NLP methodology and provides empirical evidence supporting its application in aviation safety analysis.

Technology Category

• Natural Language Processing: Safety and Robustness
• Machine Learning: Multimodal Learning
• Planning, Routing, and Scheduling: Activity and Plan Recognition

Application Category

• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Search and Retrieval-Augmented AI: Web query analysis, representation and understanding

Abstract

Safety is the main concern in the aviation industry, where even minor operational issues can lead to serious consequences. This study addresses the need for comprehensive aviation accident analysis by leveraging natural language processing (NLP) and advanced AI models to classify the phase of flight from unstructured aviation accident analysis narratives. The research aims to determine whether the phase of flight can be inferred from narratives of post-accident events using NLP. The classification performance of various models was evaluated. For single RNN-based models, LSTM achieved an accuracy of 63%, precision of 60%, and recall of 61%. BiLSTM recorded an accuracy of 64%, precision of 63%, and recall of 64%. GRU exhibited balanced performance with an accuracy and recall of 60% and a precision of 63%. Joint RNN-based models further enhanced predictive capabilities. GRU-LSTM, LSTM-BiLSTM, and GRU-BiLSTM demonstrated accuracy rates of 62%, 67%, and 60%, respectively, showcasing the benefits of combining these architectures. To provide a comprehensive overview of model performance, single and combined models were compared regarding the various metrics. These results underscore the models' capacity to classify the phase of flight from raw text narratives, equipping aviation industry stakeholders with valuable insights for proactive decision-making. Therefore, this research signifies a substantial advancement in applying NLP and deep learning models to enhance aviation safety.

Problem

• Flight Safety
• Machine Learning
• Aviation Accident Analysis

Innovation

• Deep Learning Models
• Natural Language Processing
• Aviation Safety Analysis