Cabin Safety Incident Reports for Model Training, 2004-2024

8,280 global cabin abnormal event reports from 2004 to 2024 were used to train a hybrid CNN-LSTM-Attention model for incident classification. The model achieved 95.01% accuracy and a 94.17% F1 score on the test set. The dataset was authored by Lianbin Zhou and last updated in May 2026.

Use Cases

• Benchmarking text classification models based on the reported 95.01% accuracy of the CNN-LSTM-Attention model.
• Training models for intelligent risk assessment based on the described 'text feature–risk mechanism-hazard level' mapping system.
• Developing interpretable risk quantification tools based on the framework's focus on extracting deep semantic features and key risk patterns.

Strengths

• Contains 8,280 incident reports, providing a substantial corpus for model training.
• Covers a 20-year time range from 2004 to 2024, offering longitudinal perspective.
• The associated model achieved a high test accuracy of 95.01%, suggesting the data supports effective model development.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment for specific modeling tasks.
• The dataset is very small at 5.5 KB, indicating limited scope.

Provenance

Source

figshare

Collection Method

Likely collected from global cabin abnormal event reporting systems.

Time Range

2004-2024

Freshness

Last updated 2026-05-27 17:39:59; freshness should be verified.

Geography

Global