A Comparative Application on Clustering of Mixed-type Data Sets with kamila, k-means, k-medoids and k-prototypes Algorithms
Abstract
Cluster Analysis is one of the crucial tools which is being used in many areas of scientific researches. As known, there are many algorithms for performing Cluster Analysis.
Nowadays, the main two debates relating to these algorithms are; which one to use for mixedtype data sets and how to decide selecting the best number of clusters. In this study, KAMILA algorithm which is created very ambitiously and other algorithms used before KAMILA such as k-means, k-medoids and k-prototypes algorithms will be performed for clustering the values
of different scaled variables. With this aim, a data set of a grocery store in Istanbul will be analyzed. The company has stores in different districts of Istanbul and the customers have different demographic characteristics and different purchasing behaviors. The data set provided for 999 customers includes information such as; whether the customers are purchasing the product categories that are crucial for the company's profitability and how much the total price of the purchased items are. These data were subjected to clustering analysis for customer segmentation. As a result, it is observed that KAMILA algorithm can successfully identify the customers in the segment that can be named the gold segment.
Keywords
References
- AGGARWAL, Charu C (2015), Data mining: The textbook, Switzerland: Springer.
- CUI, Hongyan, Kuo ZHANG, Yajun FANG, Stanislav SOBOLEVSKY, Carlo RATTI and Berthold KP HORN (2017), "A Clustering Validity Index Based on Pairing Frequency", IEEE Access, 5, 24884-24894.
- EVERITT, Brian and Torsten HOTHORN. (2011). Cluster analysis An Introduction to Applied Multivariate Analysis with R (pp. 163-200): Springer.
- R Development Core Team (2008). R: A language and environment, STATISTICAL COMPUTING. R FOUNDATION FOR STATISTICAL COMPUTING and Austria. ISBN 3-900051-07-0 VIENNA, URL http://www.R-project.org.
- FOSS, Alex, Marianthi MARKATOU, Bonnie RAY and Aliza HECHING (2016), "A semiparametric method for clustering mixed data", Machine Learning, 105(3), 419-458.
- FOSS, Alexander H, Marianthi MARKATOU and Bonnie RAY (2018), "Distance Metrics and Clustering Methods for Mixed‐type Data", International Statistical Review.
- GAN, Guojun, Chaoqun MA and Jianhong WU (2007), Data clustering: theory, algorithms, and applications, (Vol. 20): Siam.
- GOWER, John C (1971), "A general coefficient of similarity and some of its properties", Biometrics, 857-871.
Details
Primary Language
Turkish
Subjects
-
Journal Section
Research Article
Authors
Emrah Bilgiç
0000-0002-9875-2299
Türkiye
Publication Date
November 30, 2019
Submission Date
January 2, 2019
Acceptance Date
November 1, 2019
Published in Issue
Year 1970 Volume: 20 Number: 2