FLIGHT DELAY PREDICTION USING HYBRID ENSEMBLE MACHINE LEARNING

 Authors:

R.Ruchitha Ratna Meri, T.Manisha Reddy, R.Bhanu Prakesh, K.Krishna Lahari Patnaik, Mrs. B.Priyanka Department of Information Technology, Maharaj Vijayaram Gajapathi Raj College, Vizianagaram, Andhra Pradesh,India

Abstract—This project presents a flight delay prediction system using a hybrid ensemble learning approach. It integrates heterogeneous data sources, including historical flight records and weather data, to enhance predictive performance. Advanced feature engineering techniques are employed, such as rolling statistics based on airports and airlines, along with temporal features like hour, day, and month. Weather-related variables,including temperature and precipitation,are also incorporated to capture environmental influences on flight delays. This model architecture combines multiple machine learning algorithms, including Random Forest and Artificial Neural Networks (ANN), to form an ensemble model that improves accuracy and robustness. The proposed system achieves an accuracy of 82%, demonstrating improved performance over individual models. The system achieves high predictive performance and is deployed through a web-based interface enabling real-time user interaction and highlighting its practical applicability in real-world aviation scenarios.

Keywords–Flight delay, Machine learning, ensemble model, Risk score, Flask.