Resampling and Ensemble Strategies for Churn Prediction

Serra Çelik; Seda Tolun Tayalı

Article Details

Resampling and Ensemble Strategies for Churn Prediction

Authors:

Serra Çelik | Seda Tolun Tayalı

Publisher:

Gazi University

Main Topics:

Makine Öğrenme (Diğer)

Journal:

Bilişim Teknolojileri Dergisi Vol.16 Issue 4, pp.263-273

Document Type:

Article

Publication Date:

10/2023

Document Language:

English

Document ID (Org No.):

2001302787

pISSN:

1307-9697

Keywords:

Abstract:

 Churn analysis is a customer relationship management analytics that companies implement to predict the customers who are likely to terminate doing business with them. The success of marketing efforts to retain the existing customers is possible only if probable churners are correctly specified beforehand. Therefore, having powerful models with high prediction capabilities that lead to a profit growth is crucial. The imbalanced nature of churn datasets negatively effects the classification performance of machine learning methods. This study examines resampling over- and under-sampling- and ensemble learning bagging, boosting, and stacking strategies integrated with the cross-validation procedure on imbalanced churn prediction. The experimental results, which are compared to the results of Support Vector Machines taken as the benchmark, show that ensemble methods improve the prediction performances. Also, applying over-sampling achieves a noticeable performance in comparison with the under-sampling approach.

 Churn analysis is a customer relationship management analytics that companies implement to predict the customers who are likely to terminate doing business with them. The success of marketing efforts to retain the existing customers is possible only if probable churners are correctly specified beforehand. Therefore, having powerful models with high prediction capabilities that lead to a profit growth is crucial. The imbalanced nature of churn datasets negatively effects the classification performance of machine learning methods. This study examines resampling over- and under-sampling- and ensemble learning bagging, boosting, and stacking strategies integrated with the cross-validation procedure on imbalanced churn prediction. The experimental results, which are compared to the results of Support Vector Machines taken as the benchmark, show that ensemble methods improve the prediction performances. Also, applying over-sampling achieves a noticeable performance in comparison with the under-sampling approach.

 Churn analysis is a customer relationship management analytics that companies implement to predict the customers who are likely to terminate doing business with them. The success of marketing efforts to retain the existing customers is possible only if probable churners are correctly specified beforehand. Therefore, having powerful models with high prediction capabilities that lead to a profit growth is crucial. The imbalanced nature of churn datasets negatively effects the classification performance of machine learning methods. This study examines resampling over- and under-sampling- and ensemble learning bagging, boosting, and stacking strategies integrated with the cross-validation procedure on imbalanced churn prediction. The experimental results, which are compared to the results of Support Vector Machines taken as the benchmark, show that ensemble methods improve the prediction performances. Also, applying over-sampling achieves a noticeable performance in comparison with the under-sampling approach.

 Müşteri kayıp analizi; şirketlerin, kendileriyle çalışmayı sonlandırması muhtemel müşterileri tahmin etmek için kullandığı bir müşteri ilişkileri yönetimi analitiğidir. Mevcut müşterileri elde tutmaya yönelik pazarlama çalışmalarının başarısı, ancak olası müşteri kayıplarının önceden doğru bir şekilde belirlenmesiyle mümkündür. Bu nedenle kâr artışına yol açacak, yüksek tahmin kabiliyetli, güçlü modellere sahip olmak çok önemlidir. Kayıp analizi için kullanılan veri kümelerinin dengesiz doğası, makine öğrenimi yöntemlerinin sınıflandırma performansını olumsuz etkilemektedir. Bu çalışma, dengesiz kayıp tahmini üzerinde çapraz doğrulamanın prosedürüyle entegre edilmiş yeniden örnekleme - aşırı ve az örnekleme - ve topluluk öğrenme - bagging, boosting, ve stacking - stratejilerini incelemektedir. Referans noktası olarak alınan Destek Vektör Makinelerinin sonuçlarıyla karşılaştırılan deneysel sonuçlar, topluluk yöntemlerinin tahmin performanslarını iyileştirdiğini göstermektedir. Ayrıca aşırı örneklemenin uygulanması, az örnekleme yaklaşımına kıyasla fark edilebilir bir performans artışı sağlamıştır.

Article Details

Resampling and Ensemble Strategies for Churn Prediction

Authors:

Serra Çelik | Seda Tolun Tayalı

Publisher:

Gazi University

Main Topics:

Makine Öğrenme (Diğer)

Journal:

Bilişim Teknolojileri Dergisi Vol.16 Issue 4, pp.263-273

Document Type:

Article

Publication Date:

10/2023

Document Language:

English

Document ID (Org No.):

2001302787

pISSN:

1307-9697

churn prediction

class imbalance

resampling

support vector machines

evaluation metrics

ensemble learningkayıp tahmini

sınıf dengesizliği

yeniden örnekleme

destek vektör makineleri

değerlendirme metrikleri

topluluk öğrenme

Issues

Article Details

Authors: Serra Çelik | Seda Tolun Tayalı

Publisher: Gazi University

Main Topics: Makine Öğrenme (Diğer)

Journal: Bilişim Teknolojileri Dergisi Vol.16 Issue 4, pp.263-273

Document Type: Article

Publication Date: 10/2023

Document Language: English

Document ID (Org No.): 2001302787

pISSN: 1307-9697

Issues

Authors:

Serra Çelik | Seda Tolun Tayalı

Publisher:

Gazi University

Main Topics:

Makine Öğrenme (Diğer)

Journal:

Bilişim Teknolojileri Dergisi Vol.16 Issue 4, pp.263-273

Document Type:

Article

Publication Date:

10/2023

Document Language:

English

Document ID (Org No.):

2001302787

pISSN:

1307-9697