Article Details

Abstract

With the advancement of information technology and the intensification of market competition, customer churn has become a significant factor affecting the profitability and market share of telecommunications operators. To effectively prevent customer churn, this paper proposes a method based on Chi-Square Test and Ensemble Learning, aiming to improve the accuracy and reliability of the prediction of telecommunications customer churn. Firstly, Chi-Square Test is used to identify the features that have a significant impact on customer churn. Then, prediction models are constructed by integrating two algorithms, Random Forest and XGBoost. The experiment shows that the overall prediction performance of the model in this paper is high, which can provide a reference for customer management and marketing strategies of telecommunications enterprises.

Keywords

Chi-Square Test; customer churn; prediction model; Ensemble Learning

References

1. [1] Wang Yulin Research on the prediction model of customer churn in broadcasting and television based on ensemble learning [J]. TV Technology, 2024, 48 (7): 59-66
2. [2] Liu Mei, Zheng Lijun, Duan Yongliang, etc Customer churn prediction based on PCA+GWO integrated feature selection and model stacking [J/OL]. Computer Engineering and Applications, 2024:1-16 (2024-08-26) [2024-09-09] http://kns.cnki.net/kcms/detail/11.2127.TP.20240823.1756.009.html.
3. [3] Zeng Jiayu Telecom customer churn prediction model based on Stacking ensemble algorithm [D]. Harbin: Harbin Institute of Technology, 2022
4. [4] Fang Xiangzhong Chi square distribution and chi square test [J]. Chinese Journal of Statistics, 2022 (5): 29-31
5. [5] Liu Yaxiang, Liang Tao, Liu Yuchang Research on the influencing factors of agricultural carbon emissions in Shandong Province based on random forests [J]. Agriculture and Technology, 2024, 44 (21): 90-93
6. [6] BREIMAN L. Bagging Predictors [J].Machine Learning，1996，24（2）：123-140.
7. [7] CHEN T，GUESTRIN C. XGBoost: A Scalable Tree Boosting System [C]//KDD '16: Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. Association for Computing Machinery：ACM，2016：785-794.
8. [8] Xing Mingxuan, Zhao Jinyan Prediction of car insurance claim amount combination model based on improved Boosting algorithm [J]. Technology and Innovation, 2024 (9): 1-6
9. [9] Su Wenxing, Zhang Zhenyi, Zheng Yanli, etc Research on the Risk Prediction Model of Coronary Heart Disease Based on Ensemble Learning [J]. Intelligent Computer and Application, 2022, 12 (7): 8-13, 19
10. [10] Chi Guotai, Wang Shanshan, Wang Yiran Research on Default Risk Warning of Listed Companies Based on Stacking Method [J]. Systems Engineering Theory and Practice, 2025, 45 (2): 481-502
11. [11] ABDALLAH WAGIH IBRAHIM. Telco Customer Churn [EB/OL].[2024-10-19].https://www.kaggle.com/datasets/abdallahwagih/telco-customer-churn.
12. [12] Wang Shengjie, Zhang Qinghong Research on Customer churn prediction in the telecommunications industry based on interpretable machine learning models [J]. Telecommunications Science, 2024, 40 (7): 121-133

About the authors

Zhang Heyu (2004-), female, Han ethnicity, from Shangqiu, Henan Province, undergraduate studies, research direction: data science and big data technology; Yu Shihua (2004-), female, Han ethnicity, from Meizhou, Guangdong Province, undergraduate studies, research direction: data science and big data technology; Corresponding author: Wang Wanwan (1992-), female, Han nationality, from Zhengzhou, Henan Province, senior engineer, lecturer, master's degree, research direction: data mining, computer vision.