CLUSTERING ALGORITHM UNTUK PENGELOMPOKAN PELANGGAN DALAM BIDANG USAHA SERVER RELOAD

Abstract

Server pulsa became a modern trading technology community in this era and the
perpetrators are mostly small entrepreneurs. In the implementation there are several
problems. One of the problems of the pulse server is the promo strategy. The promo
strategy in the pulse server business generally uses the classical method by
distributing flyers to anyone without considering the target. The results obtained are
also not effective. In this paper we will discuss one of the ideas on how to use the
existing data for the promo strategy. Every day has a lot of transaction data
generated in the pulse server. These data mostly become passive data that has no
benefits for business owners, while the existence of data is very much and
accumulate in the area of data backup. Based on two things that have been
submitted, it is necessary to do a preliminary study about the importance of grouping
customer segment using clustering algorithm technique. The suitable technique is
chosen because it is considering the clustering characteristics that can bring up
previously unknown groups of attributes so that it can become knowledge for the
data owner in using it. As an illustration of the design, the attributes involved are:
customer, transaction number, and domicile to establish cluster groups.