APPLICATION OF DATA MINING FOR EVALUATION OF BEST SELLING INTERNET PACKAGE USING K-MEANS CLUSTERING
Abstract
Data Mining is a process to obtain useful information from a database warehouse in the form of knowledge. Data transaction history of sales can be informative for a business decision. One of the largest telecommunications companies in Indonesia is experiencing problems selling its products in Jambi city. The sales trend for the products has decreased last year. This is due to competitor penetration which increasingly threatens the product’s existence as the company's best-selling product. Several previous studies used analysis based on clustering data to form information. Therefore, this study aims to process sales data into several clusters using the K-means clustering algorithm. The results of the clustering form 5 package categories, namely very in demand (1 data), in demand (1 data), moderate in demand (1 data), less in demand (1 data) and not-in demand (74 data). The grouping of these 5 clusters is based on evaluation by Davies Bouldin Index which yields a value of 0.035. The results of this clustering can be useful for company management to develop marketing strategies in the future.