Empirical Evaluation of Gated Recurrent Neural Network Architectures in Aviation Delay Prediction
Authors:
- 14 October 2020
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1-7
Abstract
The aviation industry is one of the largest industries in the transportation sector. One of the major problems faced in this industry is the prevalence of delays. These delays not only cause dissatisfaction among the passengers but also huge losses for airlines. In this study, an approach using gated recurrent neural networks is explored for the purpose of predicting delays encountered in the aviation industry. The proposed approach uses Long Short-Term Memory(LSTM) architectures such as vanilla LSTM and Bi-directional LSTM to empirically evaluate its efficiency in predicting the delay for a future flight. While previous works have explored this approach for classifying the delay status as either delay or no-delay, we take this a step further by predicting the amount of delay that will be experienced. The trained models were tested on an unseen test data set and were able to predict the delay with an error value of 33 minutes.Keywords
This publication has 10 references indexed in Scilit:
- Iterative machine and deep learning approach for aviation delay prediction2017
- A deep learning approach to flight delay prediction2016
- Real-Time Investigation of Flight Delays Based on the Internet of Things Data2016
- LSTM: A Search Space OdysseyIEEE Transactions on Neural Networks and Learning Systems, 2016
- Logistics and Transportation Security2015
- Speech recognition with deep recurrent neural networks2013
- The Forecasting Model of Flight Delay Based On DMT-GMT ModelPhysics Procedia, 2012
- A Novel Connectionist System for Unconstrained Handwriting RecognitionIEEE Transactions on Pattern Analysis and Machine Intelligence, 2008
- Framewise phoneme classification with bidirectional LSTM and other neural network architecturesNeural Networks, 2005
- Bidirectional recurrent neural networksIEEE Transactions on Signal Processing, 1997