MACHINE LEARNING SYSTEM FOR MODELLING THE LEVEL OF GAS FLARING ON AIR QUALITY IN NIGERIA
Abstract
ABSTRACT
This study presents an advanced machine learning system designed to analyze and forecast PM2.5 air pollution levels in Nigeria, with a specific focus on the impact of gas flaring activities. The system employs recursive forecasting methodology through the Sk-forecast library, utilizing historical data from 1990-2020 for PM2.5 measurements and gas flaring metrics from 1990-2011. Through comprehensive benchmarking of tree-based regressors including XGBoost, Random Forest, LightGBM, and CatBoost, the research demonstrates that CatBoost achieves superior performance with a Mean Absolute Percentage Error of 6.38%. The system's sophisticated approach incorporates advanced feature engineering techniques, including lagged features and rolling statistics, whilst maintaining model interpretability through feature importance analysis. The implementation reveals significant correlations between gas flaring metrics and air quality degradation, providing valuable insights for environmental monitoring and policy development. This research contributes to the field by establishing a reliable framework for predicting air quality trends in Nigeria, offering potential applications for environmental management and decision-making processes.
Full-Text Access Notice
In accordance with the NERD Policy on promoting peer-reviewed publication, public access to the full text of a project, thesis or dissertation is restricted for three years, allowing the author and supervisors sufficient time to pursue peer-reviewed publication.
During this period, researchers with legitimate academic or research purposes may request authorisation directly from the author to enable NERD to release the indexed full texts of the work using the form below.