Article Details

Abstract

Flight delay causes a certain degree of resource waste to airlines, airports and passengers. Studying the problem of flight delay can provide better decision support and services for the outside world. Traditional statistical methods and Machine Learning technology have some limitations in processing multicollinear data. Therefore, tensor decomposition of flight data can effectively decompose and reduce the dimension of these multidimensional flight data, and mine the hidden patterns and associations in the data. Firstly, the visual interpretation is carried out according to the factors after Tucker decomposition, and then the CANDECOMP/PARAFAC (CP) decomposition is used for some interpretation. Finally, the CP decomposition and Tucker decomposition are compared. The study finds that under different airport modes and the same weather mode, the flight delay degree may have the same relationship, and the airport mode has a certain impact on the flight delay. In addition, under heavy rain, the takeoff and landing of airport flights may be greatly affected and prone to severe delays.

Keywords

flight delay; tensor decomposition; CANDECOMP/PARAFAC decomposition; Tucker decomposition

References

1. [1] Robinson P J. The Influence of Weather on Flight Operations at the Atlanta Hartsfield International Airport[J]. Weather Forecast，1989，4（4）：461-468.
2. [2] Hitchcock F L. The Expression of a Tensor or a Polyadic as a Sum of Products[J].Journal of Mathematics and Physics，1927，6（1）：164-189.
3. [3] Tucker L. Some Mathematical Notes on Three-Mode Factor Analysis[J].Psychometrika，1966，31（3）：279-311.
4. [4] Kolda T G，Bader B W. Tensor Decompositions and Applications[J].Society for Industrial and Applied Mathematics Review，2009，51（3）：455-500.
5. [5] Andrzej C，Mandic D，Lathauwer L D，et al. Tensor Decompositions for Signal Processing Applications From Two way to Multiway Component Analysis[J]. IEEE Signal Processing Magazine，2015，32（2）：145-163.
6. [6] Fu X，Vervliet N，Lathauwer L D，et al. Computing Large-Scale Matrix and Tensor Decomposition With Structured Factors:A Unified Nonconvex Optimization Perspective[J].IEEE Signal Processing Magazine，2020，37（5）：78-94.
7. [7] McNeil M，Bogdanov P. Multi-Dictionary Tensor Decomposition[C]//2023 IEEE International Conference on Data Mining (ICDM).Shanghai：IEEE，2023：1217-1222.
8. [8] Lee J，Han D，Kwon O K，et al. GPUTucker: Large-Scale GPU-Based Tucker Decomposition Using Tensor Partitioning[J/OL].Expert Systems With Applications，2024，237（PA）：121445[2025-11-07].https://www.sciencedirect.com/science/article/pii/S0957417423019474.
9. [9] Zhu Z，Sun L，Chen X Q，et al. Integrating Probabilistic Tensor Factorization with Bayesian Supervised Learning for Dynamic Ridesharing Pattern Analysis[J/OL]. Transportation Research Part C: Emerging Technologies, 2021, 125: 103010.https://doi.org/10.1016/j.trc.2020.102916.
10. [10] Sun L J，Axhausen K W. Understanding Urban Mobility Patterns with a Probabilistic Tensor Factorization Framework[J]. Transportation Research Part B: Methodological，2016，91：511-524.

About the authors

Chen Wei (1998-), male, Han ethnicity, from Yongzhou, Hunan, lecturer, master's student, research interests include mathematical statistics and machine learning.