Exploration of Evolution of Airport Ground Delay Programs

Created by Kexin Ren at University of Alberta, Canada in 2016 - 2017.

Introduction

Tool: R Techniques: Unsupervised learning (Autoencoders), clustering, statistical Analysis (Configural Frequency Analysis, CFA)

Description: explore the patterns of massive evolved air program (GDP) data

• preprocessed the discrete data into evolving forms
• visualized the parameters and time stamps of the programs
• reduced dimension (598 dim to 2 dim) using autoencoders
• clustered the self-created images
• analyzed the clusters using stat methods

Citation

If you're using this code in a publication, please cite our papers.

@article{ren2017exploration,
      title={Exploration of the Evolution of Airport Ground Delay Programs},
      author={Ren, Kexin and Kim, Amy M and Kuhn, Kenneth},
      journal={Transportation Research Record},
      pages={0361198118782272},
      year={2017},
      publisher={SAGE Publications Sage CA: Los Angeles, CA}
    }