Air Cargo Capacity Forecasting using Deep Learning Techniques

Authors

  • Krishna Kumar M, Janaki Meena M

Abstract

The international air cargo transportation sector is an important means of goods exchanges that are crucial to meet the global supply chain and market demands. The sector is always affected by a lot of uncertainties and hence the predictive modeling of the cargo capacity is the need of the hour. An accurate forecast of the capacity helps in efficient scheduling of the flights, its unit load devices (ULD) utilization, and shipment routings. In this paper, we built predictive models to forecast the cargo capacity of the passenger air crafts using state-of-the-art deep learning methods. We have done modeling using the basic statistical analysis of the data, conventional machine learning, and deep learning algorithms. The input data used for the modeling is composed of different category namely discrete, continuous, and time-series data. We made a thorough analysis of the effects of these data categories in the prediction task. Based on the study, we have developed the optimal deep learning architecture that is suited for cargo capacity forecast problems. The hybrid architecture of LSTM and dense layers had superior results compared to other approaches.

Published

2020-02-29

Issue

Section

Articles