Aeolus
收藏资源简介:
Aeolus是一个大规模的多模态航班延误数据集,旨在推进航班延误预测研究,并支持表格数据基础模型的发展。该数据集提供了三种对齐的模态:(i)一个表格数据集,包含超过5000万次航班的丰富操作、气象和机场级别的特征;(ii)一个航班链模块,通过链接连续的航班环节来模拟延误传播,捕捉上游和下游的依赖关系;(iii)一个航班网络图,编码共享的飞机、机组人员和机场资源连接,使跨航班关系推理成为可能。数据集经过精心构建,具有时间划分、综合特征和严格的泄露预防,以支持现实和可重复的机器学习评估。Aeolus支持广泛的任务,包括回归、分类、时间结构建模和图学习,作为表格、序列和图模态的统一基准。我们发布了基线实验和预处理工具,以促进采用。Aeolus填补了特定领域建模和通用结构化数据研究的关键差距。
Aeolus is a large-scale multimodal flight delay dataset designed to advance flight delay prediction research and support the development of tabular data foundation models. This dataset provides three aligned modalities: (i) a tabular dataset containing rich operational, meteorological, and airport-level features for over 50 million flights; (ii) a flight chain module that simulates delay propagation by linking consecutive flight segments to capture upstream and downstream dependencies; (iii) a flight network graph that encodes shared aircraft, crew, and airport resource connections to enable cross-flight relational reasoning. The dataset is meticulously constructed with temporal partitioning, comprehensive features, and strict leakage prevention to support realistic and reproducible machine learning evaluations. Aeolus supports a wide range of tasks, including regression, classification, temporal structure modeling, and graph learning, serving as a unified benchmark for tabular, sequential, and graph modalities. We have released baseline experiments and preprocessing tools to facilitate adoption. Aeolus fills a critical gap in both domain-specific modeling and general structured data research.

- 1Aeolus: A Multi-structural Flight Delay Dataset四川大学, 香港科技大学(广州) · 2025年



