Detection of Difference between Data based on Cluster Scale and Its Application to Market Data

Abstract

In this paper, we examine the effectiveness of classification using the cluster scale for the classification problem of the customer base in the marketing field. Specifically, the cluster scale's effectiveness was shown by comparing and verifying the misclassification rate based on the normal Euclidean distance and the method using the cluster scale applied in this study. This research can contribute to customer analysis by taking an appropriate marketing approach to a customer segment by extracting customer characteristics from large-scale customer data.