Flight Delay Prediction based on Aviation Big Data and Machine Learning
Precise prediction of delays is essential if the airline industry is to be more efficient. Recent research focused on the use of machine learning to anticipate delays in flight. Most of the forecast techniques are carried out on a particular route or airport. This article examines a wider range of variables that may affect flight delay and compares several learning models for machine-based prediction tasks. Automatic dependent surveillance (ADS-B) communications are received, pre-processed, and included with other information such as meteorological conditions, flight schedule, and airport information in order to create a dataset for the proposed system. The prediction tasks developed include several categorization tasks and a regression. Experimental findings indicate that the long-term short-term memory (LSTM) can handle data collected from the air series, however in our small dataset there is a fit issue. The suggested random forestry model may achieve better prediction precision (90.2 percent for binary classification) compared to earlier methods and avoid the issue of over fitting.