Unsupervised aided investigation on the associations between municipal solid waste characteristics and socio-economic conditions

Abstract

Better municipal solid waste (MSW) management can help to address environmental concerns and supports economic and social development. Because MSW characteristics can change over time, management strategies should also evolve and be applied correspondingly. However, many previous studies have focused on MSW characterization or investigated specific management strategies for a target MSW. Few studies have assessed the spatial variations of MSW characteristics and socio-economic (SE) conditions as well as their associations. This study evaluated the feasibility of using an integrated unsupervised method (cluster analysis and cross-tabulation analysis) to explore these topics for MSW management. Results suggest that the integrated method can successfully help to reveal key information. Seven jointed MSW-SE scenarios were investigated based on 259 individual observations of Taiwan. Associations between MSW compositions and SE conditions were identified statistically significant for four MSW-SE scenarios. In general, the general SE type (SE₁) is very likely to generate high food wastes and other combustible, low paper, wood, and rubber wastes (MSW₁). The small island SE type (SE₃) is more likely to produce high paper and low wood, rubber, textile, and other noncombustible wastes (MSW₂). Overall, the method applied in this study could help to reveal statistical associations between MSW and SE, which can help decision-makers comprehend underlying facts and develop effective management strategies.

Keywords: Cluster analysis; Cross-tabulation analysis; Municipal solid waste; Socio-economic; Spatial variation; Unsupervised learning.