Vol. 12(3), pp. 140-147. , 2024.

ISSN: 2384-6151

https://doi.org/10.14662/ijebm2024200

 

 

Full Length Research

 

Exploration of Data mining technique for the Development of Flight Delay Prediction Model for local flights in Nigeria

 

1Ogbogbo Gabriel and 2Enomate A.

 

1&2 Department of Statistics, Delta State Polytechnic, Otefe-Oghara; 2 Department of Electrical Engineering, Delta State Polytechnic, Otefe-Oghara  

 

 

 

Accepted 22 May 2024

Abstract

 

This study focuses on developing and evaluating flight delay prediction models tailored for local flights at Warri and Benin airports in Nigeria. The objective is to enhance operational efficiency and passenger satisfaction by accurately forecasting flight delays. Flight data from the Nigerian aviation database and meteorological data were utilized over a two-year period, employing rigorous data mining techniques for analysis. Data preprocessing involved cleaning, integrating, and selecting features to prepare datasets for modeling. Various predictive models, including linear regression, decision trees, random forest, support vector machines (SVM), and ensemble methods (random forest and gradient boosting), were developed. Cross-validation techniques were applied to assess model performance. Key findings include the ensemble model combining random forest and gradient boosting consistently demonstrating superior performance with an average mean absolute error (MAE) of 8.2 minutes. In comparison, individual models such as linear regression (13.0 minutes), decision tree (10.8 minutes), SVM (11.6 minutes), and random forest (8.9 minutes) achieved higher MAE values. Flight data analysis revealed specific delay patterns and arrival times crucial for model development and validation. Cluster analysis categorized days into low, medium, and high delay clusters, providing insights for proactive delay management strategies. Evaluation of empirical cumulative distribution function (ECDF) of prediction errors demonstrated varying accuracies across different delay levels, emphasizing the models' capabilities and areas for further refinement The study underscores the effectiveness of ensemble models for improving flight delay predictions in Nigerian local aviation. These findings contribute to enhancing decision-making processes in airport operations and optimizing resource allocation, ultimately benefiting passenger experience and service reliability.

 

Keywords: Flight delay prediction, Data mining, Ensemble models, Nigerian aviation, Cluster analysis, Empirical cumulative distribution function

 

Cite This Article As: Ogbogbo, G., Enomate, A. (2024). Exploration of Data mining technique for the Development of Flight Delay Prediction Model for local flights in Nigeria. Inter. J. Econ. Bus. Manage. Vol. 12(3), pp. 140-147.