A grid search optimized extreme learning machine approach for customer churn prediction

  • Fatma Önay Koçoğlu Muğla Sıtkı Koçman University
  • Tuncay Özcan İstanbul Technical University, Turkey


Customers' behaviors such as tendencies, loyalty status, satisfaction criteria show an alteration day by day due to the changing world. So, these behavior changes should be analyzed very well in every step of the decision-making process. Customer churn analysis is the determination of customers who tend to leave by analyzing the customer data with various methods before this situation occurs. Customer churn analysis is very important to take the proper steps to minimize customer losses. In this study, a new approach based Extreme Learning Machine (ELM) has been used to solve customer churn prediction problem. It is aimed to investigate the parameters of the algorithm that produce the best solution with grid search optimization. Also, a modified accuracy calculation approach has been presented. The churn data set obtained from the UCI Machine Learning Repository has been used to determine the effectiveness of the ELM. Naive Bayes (NB), k-Nearest Neighbor (kNN), and Support Vector Machine (SVM) methods are selected for performance comparison of the model. With a value of 93.1%, the best accuracy measure has been obtained with ELM. Due to the low number of parameters to be determined and performance evaluation measures that compete with other models’ results, it can be said that the ELM is highly effective and interesting in the solution of the problem.

Author Biographies

Fatma Önay Koçoğlu, Muğla Sıtkı Koçman University

Dr. Fatma Önay Koçoğlu graduated from Istanbul University Faculty of Science Department of Mathematics and also from Sakarya University Faculty of Engineering Industrial Engineering Department. She received her master's degree with the thesis titled "Comparison of Data Discretization Methods in Data Mining and an Application" from Istanbul University Informatics Department. She studied on her master thesis at Université de Technologie de Compiégne (France) Computer Engineering Department for 6 months. Koçoğlu received her Ph.D. in 2017 with her thesis "Analytical Approaches to Solving the Customer Churn Analysis Problem" from İstanbul University Informatics Department. She is continuing her second doctorate education in Istanbul University-Cerrahpaşa Department of Industrial Engineering. Her primary research interests are in data mining, machine learning, artificial intelligence, and optimization. She worked as a Research Assistant at Istanbul University (IU) Department of Informatics between 2010-2021. Koçoğlu worked as a visiting researcher at Sorbonne University LIP6. She is currently a faculty member at Muğla Sıtkı Koçman University, Faculty of Engineering, Department of Software Engineering. She has many scientific studies published in national and international fields. She worked as a researcher in projects supported by The Scientific and Technological Research Council of Turkey (TÜBİTAK) and other Higher Education Institutions. She also worked as an instructor in the scientific training programs organized by Higher Education Institutions and Non-Governmental Organizations. She speaks fluent English and basic French. 

Tuncay Özcan, İstanbul Technical University, Turkey

Dr. Tuncay Özcan completed his undergraduate education in Industrial Engineering in 2002, his graduate education in 2005, and his doctorate in 2011 with his thesis on "An analytic approach based on data mining to shelf space management in retail industry". Between 2005 and 2012, he worked as Planning and Stock Control Manager and R&D Manager in the retail sector. In 2012-2019, he worked as an Assistant Professor in the Industrial Engineering Department of the Faculty of Engineering at Istanbul University. Dr. Özcan is currently an Associate Professor at the Department of Management Engineering at the Faculty of Management at Istanbul Technical University. He has numerous international and national publications on data mining, retail management, decision making, logistics and supply chain management, and heuristic optimization.

Computer Engineering