Penggunaan Pendekatan Deep Constrained Clustering untuk Membuat Profil Bisnis

Abstract

In the development of technology at this time, data processing has been massively carried out by carrying out several approaches and data constructions. Data min- ing is the process of extracting data from a large database so that it can be used to build multivariate data, where more and more data produces increasingly complex value variations. Identification of customers in the business sector really needs to be done as an evaluation of a business that has been run in order to continue to grow and be able to keep abreast of business developments in the same sector. In this study, clustering of customers using rail mass transportation was carried out using the deep constraint clustering approach. Deep constrained clustering com- bines equality constraints between several customers to find out which of several customers in the data set are in the same group (positive or must-link constraints) and which are not in the same group (negative or cannot-link constraints). The result achieved is the formation of 6 clusters in the population for multivariables, which are obtained in the form of passenger clusters that have a tendency to use the rail mass transportation mode. By knowing the characteristics of passengers in each cluster, it can be considered for the services to be provided so as to be able to provide increased business services that have been carried out.