Forecasting Air Quality in Malaysian Cities Using Generalized Additive Mixed Model (GAMM)

Chin, Shi Yu (2025) Forecasting Air Quality in Malaysian Cities Using Generalized Additive Mixed Model (GAMM). Final Year Project (Bachelor), Tunku Abdul Rahman University of Management and Technology.

Text RMM_Chin Shi Yu_Full Text.pdf Restricted to Registered users only Download (2MB)

Abstract

This study aims to develop a forecasting model for predicting Air Quality Index (AQI) values in Malaysian cities using Generalized Additive Mixed Model (GAMM). The model incorporates key meteorological as well as spatial random effects to explain AQI variability across different urban regions. The methodology involves optimizing smooth functions for selected predictors and evaluating model performance through key metrics, including R- squared (R2), Root Mean Square Error (RMSE) and Mean Absolute Error (MAE). The model successfully explains over 80% of the variance in AQI values and meets benchmark thresholds with RMSE and MAE values, indicating strong predictive accuracy. Key findings include the identification of significant predictors of AQI and the overall good performance of the model across all cities. However, limitations were observed including the model's sensitivity to unaccounted events such as industrial emissions or haze. The study concludes with practical recommendations to integrate real-time data, improve spatial granularity and extend forecast durations to better support environmental policy and public health decision-making. Overall, this research contributes a practical, data-driven tool for short-term air quality forecasting and supports ongoing efforts to promote sustainable urban living. Keywords: Air Quality Index, Generalized Additive Mixed Model, forecasting, Malaysia, meteorological predictors, RMSE, MAE, R-squared.