Analisis K-Means Clustering pada Data Sepeda Motor

Abstract

K-Means is a data mining algorithm that can be used to grouping or clustering data. This research using k-means for clustering the data of motorcycle based on consumer needs. The dataset used in this research is Honda and Yamaha motorcycle which taken from the dialers in Dewantara District, Aceh. The data tested by grouping 300 data of motorcycle with different attributes into 3 clusters, which are cheap, normal, and expensive. The distribution of the data we separate it using 45 data in 15 times of test. Each test used 3 different data randomly selected on each test. To calculate the distance of each motorcycle data that have been inputted to each centroid, we used the Euclidean Distance formula. Data in this cluster system can be used as a recommendation for users in selecting the motorcycle that they interest the most. The results of the performance on each test finished in 15 times shown that  the average value of Precision by 76%, Recall by 76% and  the accuracy by 81%.