Forecasting Medical Waste Generation In Hospital Taiping, Perak Using Nonlinear Autoregressive (NAR) Neural Network Method

To assist in the management of medical waste such as storage,transportation, and disposal of the waste, forecasting the amount of medical waste generation is crucial. This study applies Nonlinear Autoregressive (NAR) Neural Network to forecast the generation of medical waste in Hospital Taiping, Perak for two years, from 2021 to 2022, based on data of medical waste generated during 2018 until 2020 on monthly basis that comprises 36 datasets. The data were randomly divided into 60% for training, 20% for validation, and 20% for testing. The results predicted that for 2021,the total medical waste generated would be 397 510.564 kg which shows an increasing trend compared to the previous year. For 2022,the result predicted 326 608.8845 kg of total medical waste generated which shows a decreasing trend from the predicted previous year. In conclusion,the NAR Neural Network method is considered reliable and may aid in the forecasting of medical waste generation for more efficient waste management plan.

I. Introduction

Since the past decades, the increasing human population has resulted in rising demand for medical and healthcare services. This scenario creates a need for more healthcare facilities and thus, generates huge quantity of medical waste. In Malaysia, the Department of Environment (DOE) categorizes medical waste as a scheduled waste that is regulated by the Environmental Quality (Scheduled Wastes) Regulations, 2005. This regulation requires appropriate handling and disposal of the medical waste. The components that make up the medical waste are human tissues, blood, body fluids, excretions, drugs, needles, and other substances in the same categories [1].