Article Details

Abstract

This paper takes the credit card consumption dataset as the analysis object and relies on Machine Learning technology as the core tool to systematically carry out customer segmentation and high-value target customer mining. Firstly, it gains in-depth insight into the intrinsic characteristics of the data through data preprocessing and optimizes the data structure by combining PCA dimensionality reduction and feature standardization methods. Secondly, it compares the clustering effects of the K-means algorithm and the hierarchical clustering algorithm and finally determines that the optimal number of clusters is 4. On this basis, it divides customers into 4 groups with distinct characteristics and further screens out 11 high-value target customers from them. The research results can provide quantitative data support for the formulation of precision marketing plans, the implementation of dynamic risk management and control, and the design of differentiated customer maintenance strategies in the credit card business, and effectively help the business achieve the dual goals of revenue growth and risk controllability.

Keywords

Machine Learning; K-means clustering; hierarchical clustering; high-value customer mining; marketing strategy formulation

References

1. [1] Zhou Zhihua Machine Learning [M]. Beijing: Tsinghua University Press, 2016
2. [2] Zeng Wenquan, Zhang Liangjun Python Data Analysis and Applications: 3rd Edition [M]. Beijing: People's Posts and Telecommunications Press, 2025
3. [3] Xia Guangwei, He Qingquan, Li Yushan Research on Financial Customer Risk Based on K-means Data Analysis Method [J]. Journal of Insurance Vocational College, 2022, 36 (5): 52-57
4. [4] Lan Jinghong, Huang Jianglin Credit Card Fraud Detection Based on Active Learning and Rotating Forest Algorithm [J/OL]. Computer Applications and Software, 2024: 1-8 [225-11-09] https://link.cnki.net/urlid/31.1260.TP.20240511.1613.004.
5. [5] Zeng Guowen, Zhong Ling, Luo Qionghua Research on Customer Grouping and Loss Warning Algorithm Based on Health [J]. Telecommunication Engineering Technology and Standardization, 2023, 36 (11): 49-55
6. [6] Shen Hongbo, the Cao army Research on Credit Card Risk and Consumer Behavior [M]. Shanghai: Fudan University Press, 2019
7. [7] Xu Hongfei, Mou Lingxing, Xu Wenjie, etc Research on Credit Scoring Card Combination Optimization Based on QUBO Model [J]. Science and Technology Innovation and Application, 2024, 14 (19): 43-46
8. [8] Chen Hua, Zhang Qian, Wang Zhaoli, etc Research on Comprehensive Evaluation of Water Quality Based on Principal Component Analysis and Cluster Analysis [J]. Science and Technology Innovation and Application, 2023, 13 (26): 88-92
9. [9] Du Junpeng, Lv Jun, Wu Jisheng, etc Research on River Water Quality Evaluation Based on Principal Component Analysis and Systematic Cluster Analysis [J]. Water Conservancy Technical Supervision, 2022 (12): 216-220
10. [10] Jiang Hongxun, Jiang Junyi, Liang Xun A review of research on credit card transaction fraud detection based on machine learning [J]. Computer Engineering and Applications, 2023, 59 (21): 1-25

Funding

Guangdong Province Higher Education Teaching Reform Project (2023JXGG05); Excellent Course Project of Digital Intelligence Integration (XJYXKC202539)

About the authors

Ge Yanna (1987-), female, Han nationality, from Zhoukou, Henan Province, lecturer, master's student, research direction: big data analysis and application, machine learning.