Telecom Churn Prediction System Based on Ensemble Learning Using Feature Grouping
Abstract
:1. Introduction
- (1)
- A new prediction system based on ensemble learning with relatively high accuracy is proposed.
- (2)
- New features derived from equidistant grouping of customer behavior features are used to improve the system performance.
2. Literature Review
3. Materials and Methods
3.1. Dataset Preparation
3.2. Proposed Method
3.2.1. New Feature Construction with Equidistant Grouping
3.2.2. Stacking Model
- (1)
- Classifiers
- (2)
- Stacking
3.2.3. Soft Voting
3.3. Evaluation Measures
4. Results
4.1. Feature Construction
4.2. Stacking Model
4.3. Soft Voting and Final Results
4.4. Comparison with Other Works
5. Conclusions
Author Contributions
Funding
Institutional Review Board Statement
Informed Consent Statement
Data Availability Statement
Conflicts of Interest
References
- García, D.L.; Nebot, À.; Vellido, A. Intelligent data analysis approaches to churn as a business problem: A survey. Knowl. Inf. Syst. 2017, 51, 719–774. [Google Scholar] [CrossRef] [Green Version]
- Borja, B.; Bernardino, C.; Alex, C.; Ricard, G.; David, M.-M. The Architecture of a Churn Prediction System Based on Stream Mining. Front. Artif. Intell. Appl. 2013, 256, 157–166. [Google Scholar] [CrossRef]
- Kotler, P.T. Marketing Management: Analysis, Planning, Implementation and Control; Prentice-Hall: London, UK, 1994. [Google Scholar]
- Ngai, E.; Xiu, L.; Chau, D. Application of data mining techniques in customer relationship management: A literature review and classifification. Expert Syst. Appl. 2009, 36, 2592–2602. [Google Scholar] [CrossRef]
- Motoda, H.; Liu, H. Feature Selection, Extraction and Construction; Communication of IICM (Institute of Information and Computing Machinery): Taipei, Taiwan, 2001; Volume 5, pp. 67–72. [Google Scholar]
- Edwards, R.A.H.; Šúri, M.; Huld, T.A.; Dallemand, J.F. GIS-Based Assessment of Cereal Straw Energy Resource in the European Union. Available online: http://citeseerx.ist.psu.edu/viewdoc/download? (accessed on 10 February 2020).
- Sharma, A.; Panigrahi, D.P.K. A Neural Network based Approach for Predicting Customer Churn in Cellular Network Services. Int. J.Comput. Appl. 2013, 27, 26–31. [Google Scholar] [CrossRef]
- Amin, A.; Al-Obeidat, F.; Shah, B.; Adnan, A.; Loo, J.; Anwar, S. Customer churn prediction in telecommunication industry using data certainty. J. Bus. Res. 2019, 94, 290–301. [Google Scholar] [CrossRef]
- Vijaya, J.; Sivasankar, E. An efficient system for customer churn prediction through particle swarm optimization based feature selection model with simulated annealing. Clust. Comput. 2019, 22, 10757–10768. [Google Scholar] [CrossRef]
- Umayaparvathi, V.; Iyakutti, K. Applications of Data Mining Techniques in Telecom Churn Prediction. Int. J. Comput. Appl. 2012, 42, 5–9. [Google Scholar] [CrossRef]
- Jahromi, A.T.; Moeini, M.; Akbari, I.; Akbarzadeh, A. A Dual-Step Multi-Algorithm Approach for Churn Prediction in Pre-Paid Telecommunications Service Providers. J. Innov. Sustain. RISUS 2010, 1, 2179–3565. [Google Scholar] [CrossRef] [Green Version]
- Zhang, Y.; Qi, J.; Shu, H.; Cao, J. A hybrid KNN-LR classifier and its application in customer churn prediction. In Proceedings of the 2007 IEEE International Conference on Systems, Man and Cybernetics, Montréal, QC, Canada, 7–10 October 2007; pp. 3265–3269. [Google Scholar] [CrossRef]
- Reichheld, F.F.; Sasser, W.E. Zero defections: Quality comes to services. Harv. Bus. Rev. 1990, 68, 105–111. [Google Scholar] [PubMed]
- Obiedat, R.; Al-kasassbeh, M.; Faris, H.; Harfoushi, O. Customer churn prediction using a hybrid genetic programming approach. Sci. Res. Essays 2013, 8, 1289–1295. [Google Scholar] [CrossRef]
- Sabbeh, S.F. Machine-Learning Techniques for Customer Retention: A Comparative Study. Int. J. Adv. Comput. Sci. Appl. 2018, 9. [Google Scholar] [CrossRef] [Green Version]
- Faris, H. A Hybrid Swarm Intelligent Neural Network Model for Customer Churn Prediction and Identifying the Influencing Factors. Information 2018, 9, 288. [Google Scholar] [CrossRef] [Green Version]
- Larose, D.T.; Larose, C.D. Discovering Knowledge in Data: An Introduction to Data Mining; John Wiley & Sons: New York, NY, USA, 2014. [Google Scholar]
- Domingos, P. A few useful things to know about machine learning. Commun. ACM 2012, 55, 78–87. [Google Scholar] [CrossRef] [Green Version]
- García-Torres, M.; Gómez-Vela, F.; Becerra-Alonso, D.; Melián-Batista, B.; Moreno-Vega, J.M. Feature grouping and selection on high-dimensional microarray data. In Proceedings of the 2015 International Workshop on Data Mining with Industrial Applications (DMIA), San Lorenzo, Paraguay, 14–16 September 2015. [Google Scholar]
- Sturges, H.A. The choice of a class interval. J. Am. Stat. Assoc. 1926, 21, 65–66. [Google Scholar] [CrossRef]
- Chen, T.; Guestrin, C. XGBoost: A Scalable Tree Boosting System. In Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, San Francisco, CA, USA, 13–17 August 2016. [Google Scholar] [CrossRef] [Green Version]
- Chatterjee, S.; Simonoff, J. Handbook of Regression Analysis with Applications in R. Logistic Regression; Wiley: Hoboken, NJ, USA, 2020; pp. 143–171. [Google Scholar] [CrossRef]
- Suzuki, J. Decision Trees. In Statistical Learning with Math and R; Springer: Singapore, 2020; pp. 147–170. [Google Scholar] [CrossRef]
- Larose, C.D.; Larose, D.T. NAÏVE BAYES CLASSIFICATION. In Data Science Using Python and R; John Wiley & Sons, Inc.: Hoboken, NJ, USA, 2019. [Google Scholar] [CrossRef]
- Ting, K.M.; Witten, I.H. Issues in Stacked Generalization. J. Artif. Intell. Res. 1999, 10, 271–289. [Google Scholar] [CrossRef] [Green Version]
- Zhou, Z. Ensemble Methods: Foundation Sand Algorithms; CRC Press: Boca Raton, FL, USA, 2012; ISBN 978-1-439-830031. [Google Scholar]
Feature Name | Description | Object |
---|---|---|
State | Customer State | Object |
Account length | Account used days | Int64 |
Area code | Phone area code | Int64 |
phone number | Customer phone number | Object |
International plan | Whether the customer starts international business | Object |
Voice mail plan | Whether the customer starts the voice mail service | Object |
Number vmail messages | Number of customer vmail messages | Int64 |
Total day minutes | Total minutes of talk during the day | Float64 |
Total day calls | Number of calls in the day | Int64 |
Total day charge | Call charges during the day | Float64 |
Total eve minutes | Total minutes of talk last night | Float64 |
Total eve calls | Number of calls last night | Int64 |
Total eve charge | Charges for calls last night | Float64 |
Total night minutes | Night total call minutes | Float64 |
Total night calls | Total number of calls in the evening | Int64 |
Total night charge | Total charge for calls at night | Float64 |
Total intl minutes | Total minutes of international business calls | Float64 |
Total intl calls | Total number of international business calls | Int64 |
Total intl charge | Total charges for international business calls | Float64 |
Customer service calls | The number of calls for customer service | Int64 |
Churn | Is the customer churn | Bool |
Feature | Value | K |
---|---|---|
Number vmail messages | 0–51 | 6 |
Total day minutes | 0–350.8 | 10 |
Total day calls | 0–165 | 8 |
Total day charge | 0–59.64 | 6 |
Total night minutes | 0–363.7 | 10 |
Total night calls | 0–170 | 8 |
Total night charge | 0–30.91 | 7 |
Total eve minutes | 0–395 | 10 |
Total eve calls | 0–175 | 8 |
Total eve charge | 0–17.77 | 5 |
Total intl minutes | 0–20 | 5 |
Total intl calls | 0–20 | 5 |
Instance (Index) | Original Feature | New Feature |
---|---|---|
1 | 110 | 6 |
2 | 123 | 7 |
3 | 114 | 6 |
4 | 71 | 4 |
5 | 113 | 6 |
6 | 98 | 5 |
7 | 88 | 5 |
8 | 79 | 4 |
... | ||
3333 | 113 | 6 |
Model | Accuracy | Precision | Recall | F1-Score | Time(S) | |||||
---|---|---|---|---|---|---|---|---|---|---|
ODS | NDS | ODS | NDS | ODS | NDS | ODS | NDS | ODS | NDS | |
LR | 0.8482 | 0.8586 | 0.7978 | 0.8412 | 0.8576 | 0.8745 | 0.8054 | 0.8412 | 0.0327 | 0.0606 |
DT | 0.8513 | 0.8605 | 0.8176 | 0.8487 | 0.8029 | 0.8078 | 0.8173 | 0.8287 | 0.0423 | 0.0463 |
NBC | 0.8514 | 0.8575 | 0.8445 | 0.8773 | 0.8573 | 0.8631 | 0.8545 | 0.8664 | 0.0048 | 0.0053 |
XGB | 0.9543 | 0.9554 | 0.9467 | 0.9547 | 0.9528 | 0.9548 | 0.9523 | 0.9554 | 0.5853 | 0.7207 |
Model | Accuracy | |
---|---|---|
Before Stacking | After Stacking | |
LR | 0.8586 | 0.9585 |
DT | 0.8605 | 0.9560 |
NBC | 0.8575 | 0.9535 |
Model | Accuracy | |
---|---|---|
ODS | NDS | |
LR | 0.9463 | 0.9585 |
DT | 0.9493 | 0.9560 |
NBC | 0.9471 | 0.9535 |
Model | Accuracy | |
---|---|---|
ODS | NDS | |
Proposed model | 0.9612 | 0.9809 |
Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. |
© 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/).
Share and Cite
Xu, T.; Ma, Y.; Kim, K. Telecom Churn Prediction System Based on Ensemble Learning Using Feature Grouping. Appl. Sci. 2021, 11, 4742. https://doi.org/10.3390/app11114742
Xu T, Ma Y, Kim K. Telecom Churn Prediction System Based on Ensemble Learning Using Feature Grouping. Applied Sciences. 2021; 11(11):4742. https://doi.org/10.3390/app11114742
Chicago/Turabian StyleXu, Tianpei, Ying Ma, and Kangchul Kim. 2021. "Telecom Churn Prediction System Based on Ensemble Learning Using Feature Grouping" Applied Sciences 11, no. 11: 4742. https://doi.org/10.3390/app11114742