Flight Delay Prediction Using Random Forest Classifier

Abstract

Airlines focus on minimizing cost while ensuring on-time arrivals in their operations to avoid revenue loss. Especially network carriers with hub connections ensure that the incoming flights are on time for passenger, crew, and aircraft transfer by avoiding delays. Delay in time sums up billions of dollars in the aviation sector, predicting delay time helps in re-planning flight plans in a way to avoid delay. The existing deterministic models and real-time prediction system for delay time calculation lacks accuracy. The paper mainly focuses on using the airline arrival data and building a machine learning model (Random Forest Classifier) to predict delay time and probability. As random forest, in general, is robust, more flexible, and makes effective estimates this model will help in improving the overall performance of the system.