Penentuan Pencemaran Air Menggunakan Metode Self Organizing Maps (SOM)

Authors

  • Deni Universitas Tanjungpura Pontianak
  • Dwi Marisa Midyanti Universitas Tanjungpura Pontianak
  • Rahmi Hidayati Universitas Tanjungpura Pontianak

DOI:

https://doi.org/10.33022/ijcs.v11i1.3036

Keywords:

Pencemaran Air, Clustering, SOM, Silhouette Coefficient

Abstract

Water pollution is the entry of a component or substance into water, which causes water quality to decrease. The decrease in water quality causes the water to be unfit and affects the community's health. Therefore, an application for determining the water quality level building using the SOM algorithm with several data input parameters: Fluoride, Hardness, and Nitrite. These parameters use for data clustering in determining clean and polluted water. The SOM algorithm arranges SOM neurons based on the data input values ​​in a SOM cluster. For testing the SOM method on 114 data, the Silhouette Coefficient value was used to find the best number of clusters. Silhouette Coefficient will evaluate clustering the proximity of the distance between data. The test was carried out 27 times with variations of the experiment with the learning rate starting from 0.1 to 0.9 and the number of clusters from 2 to 4 to get the best Silhouette Coefficient value. The result of clustering the best silhouette coefficient value obtained is 0.7276473444141 with 3 clusters and the learning rate of 0.2.

Downloads

Published

30-04-2022