Browsing by Author "Abdalla, Olla Nagimeldin Fadul"

The waste management process is inherently intricate, involving a complex interplay of technical, climatic, environmental, socio-economic, and demographic factors. This study aims to propose a more adaptable approach to enhance the precision of predictions in volatile scenarios, exemplified by the challenges posed by COVID-19, addressing the limitations of conventional methods' inflexibility. In this analysis, the data on municipal waste collection were collected from the Solid Waste Management and Public Cleaning Corporation (SWCorp) and the National Solid Waste Management Department (JPSPN). Data spans from January 2012 to December 2021 and aims to assess the performance of municipal waste management systems in eight states of Peninsular Malaysia during the pandemic utilizing "Integrated Wasteaware Benchmark Indicators". Moreover, the study utilized a predictive approach using artificial intelligence to forecast waste generation trends up to 2030. The study compares multiple models including artificial neural network (ANN), Gradient boosting (GB), Support vector machine (SVM), Autoregressive Integrated Moving Average (ARIMA), and Vector AutoRegressive (VAR). Each model’s accuracy is evaluated using two evaluation metrics, Mean Absolute Error (MAE) and coefficient of determination (R2 ). Additionally, this study also studies the impact of data preprocessing on the AI model's performance. The hybrid model of gradient boosting and ARIMA algorithm using preprocessed data was found to have the lowest MAE and the highest R2 values of 0.341 and 0.954 respectively. The study's findings demonstrate that the developed model can provide targeted parameter predictions and recommendations for more effective strategies for waste management in the future.