Clustering Babyshop at Marketplace X with Cluster Ensemble based on Squeezer Algorithm

Abstract

Marketplace is one of the most popular digital business in Indonesia. One of the category that grow in Marketplace X is shops that sell baby equipment or better known as babyshop. In order to provide the best service and keep the credibility of babyshop specialty shops, important to do qualitry monitoring on of them through clustering. Clustering based on store reputation assessment indicators consisting of variables that are categorical and numerical in scale. This study aims to classify babyshop on Marketplace X based on the characteristics of the store using cluster analysis with cluster ensemble based mixed data clustering (CEBMDC) based on the weigthed squeezer algorithm. This study use the stream data from 218 babyshop at Marketplace X which consists of service factors, reputation level, type and location of the babyshop. The optimal cluster results was into three clusters in which cluster one consists of 21% babyshop, cluster two 48% babyshop, and 31% babyshop at cluster three. The first cluster is a cluster with the tendency of the babyshop to be classified as good, the cluster two tend to have a normal (neutral) reputation, while members of cluster three has a for poor reputation.