Penerapan Algoritma K-Means Untuk Clustering Harga Rumah Di Bandung

Abstract

The need for shelter is one of the fundamental aspects of daily life for humans. A house serves not only as a place to seek protection and rest but also as a venue for socializing with family. One of the factors influencing the decision in choosing a house is its price. House prices vary in each region, depending on factors such as location and other attributes. In major cities like Bandung, house prices differ based on their categories. However, many people still find it challenging to determine the value and discern whether a house is classified as affordable or expensive. Hence, there is a need for a clustering process of house prices in Bandung to aid in comprehending and categorizing house prices based on attributes such as the house price, total building area, and total land area. To understand and analyze the patterns of house prices in Bandung, this study utilizes the K-Means method to cluster the house price data into several groups based on their similarity in attributes. Additionally, the research aims to determine the optimal number of clusters through the cluster validation process using the silhouette index. The findings show that when using n_cluster=2, a silhouette score of 0.8870 is obtained, and with n_cluster=3, the silhouette score is 0.8009. These results indicate that clustering with n_cluster=2 and n_cluster=3 both exhibit strong interpretative structures. Thus, the clustering of house prices in Bandung can be effectively grouped into 2 clusters, as evidenced by the higher silhouette score obtained with n_cluster=2, approaching 1 compared to n_cluster=3.