@thesis{thesis,
		author={Asri  Yessy and Kuswardani  Dwina and Wullur  Ferdinand Hendrik},
		title ={Perbandingan Metode K-Medoids Dan K-Means Dalam Pengelompokan Data Load Profile Pelanggan AMR (AUTOMATIC METER READING) PT. PLN (Persero) Distribusi Jakarta Raya},
		year={2020},
		url={http://156.67.221.169/4073/},
		abstract={The AMR (Automatic Meter Reading) system implemented by PT. PLN (Persero) is used in order to detect losses (loss of electrical energy). Non-technical shrinkage is a type of shrinkage that has a major role in electrical power losses. By comparing the K-Medoids and K-Means Clustering methods, it can help to see how optimal the clusters formed in the customer load profile data grouping and the Davies-Bouldin Index method are used to determine which cluster sets are optimal for grouping, the final result is a diagram that can be used. used as a reference for the power usage status of the customer. In the K-Medoids process, itself uses 103 (one hundred and three) training data of AMR customer load profiles and gets a cluster set of 4 as the most optimal cluster. The results of this study were found from the comparison of the Davies-Bouldin Index value and the accuracy level of data grouping using the Confusion Matrix, the K-Means DBI value was 0.893 while the K-Medoids was 1.991, but in the accuracy of grouping, the K-Medoids data was superior to K-Means with a K value. -Medoids 78.59% and K-Means value 60%.}
		}