Time series analysis for waste quantities prediction in a north African city

Abstract

Municipal Solid Waste (MSW) Generation is an important variable in waste management and used to support decision makers and planners to do short and medium term management strategies. The main objective of this paper is to assist decision makers in this process using time series analysis in a Moroccan city with a high urbanization and economic growth. For this, we study, select and evaluate waste production model using Box-Jenkins Methodology. The accurate SARIMA (Seasonal Auto-Regressive, Integrated, Moving Average) model is selected for monthly waste generation in Kenitra, Morocco, based on data from January 2005 to December 2015. The model obtained shows the presence of seasonality in monthly waste generation in this city. This model also conforms that monthly production of waste in the city of Kenitra, Morocco, depends on that of the previous month but also on the same month of the previous season. Accuracy of predictions is measured by Mean Absolute Percentage Errors (MAPE) which is equal to 6% in this case of study. Thus, time series analysis found to be adequate to predict waste generation in this north African city.