PENGELOMPOKAN KABUPATEN/KOTA DI JAWA TENGAH MENURUT KUALITAS UDARA AMBIEN MENGGUNAKAN ANALISIS K-HARMONIC MEAN CLUSTER (Studi Kasus: Kualitas Udara Ambien pada Kawasan Pemukiman di Jawa Tengah Tahun 2015)

Abstract

Ambient air is the open air which is vital to the survival of every living being on earth. The activities of people have directly impacted the ambient air quality which can be seen by the result of the ambient air quality measuring. The measuring can be done in residential areas toward some substances which parameters have been measured: Nitrogen Dioxide (NO2), Carbon Monoxide (CO), Sulfur Dioxide (SO2), Oxide (O3), Hydrocarbon (HC), dan Total Suspended Particulate /Dust (TSP). To figure out the characteristic similarity between 35 Districts in Central Java, we need to group them based on the substances’ parameter to help the government run the ambient air pollutant reducing program in each clusters. The method used is K-Harmonic Means (KHM) Cluster. KHM is an algorithm invented to complete the K-Means. In this method, the centroid is updated by put the harmonic average from all points into each centroid clusters. There would be two, three, four, five, or six clusters.We need to use the ratio of variability within clusters and ratio of variability between clusters to get a relevant count of how many clusters are. The final result shows that five clusters is the relevant amount of cluster. In the first cluster there is 1 District with the highest levels of TSP, O3, SO2 and NO2. In the second cluster there are 2 Districts with the highest level of CO and the level of NO2 as the lowest. In the third cluster there are 6 Districts with the highest CO level and the lowest O3 level. In the fourth cluster there are 21 Districts with the highest HC level, but the levels of CO, TSP, and SO2 in this cluster are lower than in the third cluster. In the fifth cluster there is 1 District with the highest level of HC but the CO, SO2, and TSP levels are the lowest in this cluster. Keywords: ambient air, K-Harmonic Means Cluster, K-Means, centroid, cluster