Komparasi Metode K-Nearest Neighbors (K-NN) Dengan Support Vector Machine (SVM) Untuk Klasifikasi Status Kualitas Air

Icha Gusti Vidiastanta (Fakultas Ilmu Komputer, Universitas Brawijaya)
Nurul Hidayat (Fakultas Ilmu Komputer, Universitas Brawijaya)
Ratih Kartika Dewi (Fakultas Ilmu Komputer, Universitas Brawijaya)

Jurnal Pengembangan Teknologi Informasi dan Ilmu Komputer
Vol 4 No 1 (2020): Januari 2020

Water quality status classification for the community is divided into 2 classes namely those that meet the standards and do not meet the standards for consumption. The field of object classification research has been carried out, making it possible to create technology in the field of object classification with high accuracy. There are many classification methods, in this study discussing the comparison between K-Nearest-Neighbors (KNN) algorithm and Support vector machine (SVM). Research on the variables in the KNN and SVM algorithm to determine the best variable in classification. Testing is done by the K-Fold method with a value of K = 5 on a dataset of water quality status. Tests carried out to get the optimal parameter value KNN with K = 7 and SVM with value of the maximum iteration value = 300, = 10−12, 𝜎 = 0.07, 𝜆 = 3, 𝛾 = 1.7, and 𝐶 = 1. This research resulted in an accuracy of KNN of 88.94% and SVM of 87.71%. It was observed that the K-Nearest-Neighbors (KNN) algorithm had higher accuracy than the Support vector machine (SVM) algorithm.