Cluster Analysis for Optimal Infectious Waste Management Using Data Mining Techniques

Abstract

Infectious waste management is a critical issue affecting the environment, public health, and operational costs, particularly during the COVID-19 pandemic, which significantly increased infectious waste volumes. This analytical research aimed to 1) to study the attributes of customers using infectious waste collection services, and 2) analyze customer data clustering using K-Means Clustering method. The study utilized data from 540 customers, comprising 37 large healthcare facilities and 503 small healthcare facilities of New Chok Amnuay Chiang Mai Company Limited. The analysis was conducted through Orange Data Mining software, considering variables including infectious waste quantity (0.5-12 kg/month), distance (0.3-15 km), and collection frequency. The results revealed three distinct customer clusters: Cluster 1 (85.66%) characterized by low waste quantity (average 1.2 kg/month), short distance (average 1.8 km), and moderate collection frequency; Cluster 2 (11.43%) with high waste quantity (average 8.5 kg/month), moderate distance (average 4.2 km), and low collection frequency; and Cluster 3 (2.91%) with moderate waste quantity (average 3.7 kg/month), long distance (average 8.5 km), and high collection frequency. The clustering showed high efficiency with a Silhouette Score of 0.873. The comparison of operational efficiency before and after implementation showed improvements in four main areas: time management, vehicle capacity utilization, collection frequency, and area management. These findings can be applied to optimize collection routes, design cluster-specific service packages, and improve resource management, leading to reduced operational costs and enhanced service quality. However, future studies should consider additional factors such as infectious waste types and GIS integration for more comprehensive and sustainable development.