LGAIJan 25, 2022

Model Generalization in Arrival Runway Occupancy Time Prediction by Feature Equivalences

arXiv:2201.11654v1
Originality Synthesis-oriented
AI Analysis

This addresses a research gap in airport operations by enabling more generalizable models for real-time runway occupancy prediction, though it is incremental as it builds on existing tree-based methods.

The paper tackled the problem of generalizing real-time arrival runway occupancy time prediction across multiple airports by substituting categorical features with numerical equivalences, showing that models trained on numerical features perform at least as well as those on categorical features and can predict on unseen airport data.

General real-time runway occupancy time prediction modelling for multiple airports is a current research gap. An attempt to generalize a real-time prediction model for Arrival Runway Occupancy Time (AROT) is presented in this paper by substituting categorical features by their numerical equivalences. Three days of data, collected from Saab Sensis' Aerobahn system at three US airports, has been used for this work. Three tree-based machine learning algorithms: Decision Tree, Random Forest and Gradient Boosting are used to assess the generalizability of the model using numerical equivalent features. We have shown that the model trained on numerical equivalent features not only have performances at least on par with models trained on categorical features but also can make predictions on unseen data from other airports.

Foundations

The foundational work for this paper's niche, ranked by how specifically the neighbourhood builds on it — not by global fame.

Your Notes