Implementation of the waste volume clustering method at company "x" to reduce the amount of waste using the k-medoids algorithm

Abstract

Garbage/waste is the residue of human daily activities or comes from natural processes in solid form. Some of the causes that affect the environment are the problems of managing and disposing of waste/household waste. Company "X" is a company engaged in waste transportation services and non-hazardous waste management for business sectors such as apartments, offices, hospitals, and hotels. In this problem, there is an increase in the cubication of household waste produced by the vendor company "X" every month. Data mining is the extraction of information. Identify hidden information from large amounts of data, leverage and promote knowledge in real-time applications. Clustering is a multidimensional statistic designed to collect similar individuals into homogeneous classes based on observations in variables. The resulting classes can be set according to various structures. This study will cluster data from the volume of levies issued by the Environmental Service from 2010 to 2021 into 3 clustering categories, namely Klassam_0 for low volume of waste generated, Klassam_1 for medium volume of waste generated, and Klassam_2 for the volume of waste generated is high. From the results of this study, there were 8 on Klassam_0, 88 on Klassam_1, and 48 on Klassam_2. The volume of organic and inorganic waste is very good because of the emphasis on the waste management process that occurs in 2021, from the results of the clustering in 1 year (12 months) in 2021 8 months or clusters that are included in klassam¬_0 which can be interpreted as volume of waste "low", and 4 months or clusters included in klassam_2 which can be interpreted as "high" volume of waste.