Comparison of XGBoost and SVM for Predicting User Behaviour Segment Shifts of QRIS Merchants Based on Dynamic RFM with SHAP Interpretation

Authors

  • Zulham Sitorus Universitas Pembangunan Panca Budi, Indonesia
  • Hanna Willa Dhany Universitas Pembangunan Panca Budi
  • Ade Iskandar Universitas Pembangunan Panca Budi

DOI:

https://doi.org/10.30865/jurikom.v13i5.9576

Keywords:

QRIS, Dynamic RFM, Segment Transfer, XGBoost, Support Vector Machine, SHAP

Abstract

With the development of QRIS, transaction traces that record user behaviour can be maintained, but retail merchants rarely utilise them to anticipate changing customer behaviour. In predicting the shift in user behaviour segments of QRIS at PMart, this study compares Extreme Gradient Boosting (XGBoost) and Support Vector Machine (SVM). Segments were formed using K-Means from RFM (Recency, Frequency, and Monetary), which were recalculated every month with a three-month window. The status of segment movement for the following month is up, stable, or down, and is predicted using dynamic RFM features and their derivatives. In the simulation data, which contains 66,158 transactions from 1,168 users, both models were validated temporally, tuned with Optuna, tested with the McNemar test and bootstrap, and explained with SHAP. Three active segments were formed, and only 63.3% of observations remained in their segments from month to month. XGBoost had a lower error (p < 0.001), but the macro-F1 difference was not significant. SVM achieved a macro-F1 of 0.612 and an MCC of 0.488. SVM is more sensitive to high-value users who are moving down the segment (recall 76.3% versus 60.7%). The explanations of both models are consistent; the main driving factors for segment migration are frequency, shopping trends, and frequency (Spearman ρ = 0.868). This framework can be used as a basis for an early warning system for customer retention in QRIS stores.

References

[1] Bank Indonesia, Peraturan Anggota Dewan Gubernur Nomor 21/18/PADG/2019 tentang Implementasi Standar Nasional Quick Response Code untuk Pembayaran. Jakarta, Indonesia: Bank Indonesia, 2019.

[2] S. Mayasari, “BI Mencatat Nilai Transaksi QRIS Capai Rp 164,48 Triliun pada Januari 2026,” Kontan.co.id, Mar. 12, 2026. [Online]. Available: https://keuangan.kontan.co.id/news/bi-mencatat-nilai-transaksi-qris-capai-rp-16448-triliun-pada-januari-2026

[3] L. Tesaloni, “Transaksi Digital Tumbuh 39,24%, Penggunaan QRIS Melonjak 75,8%,” Kontan.co.id, Oct. 23, 2026. [Online]. Available: https://keuangan.kontan.co.id/news/transaksi-digital-tumbuh-3924-penggunaan-qris-melonjak-758

[4] E. H. Muchtar et al., “Quick response code Indonesia standard (QRIS) E-payment adoption: customers perspective,” Cogent Business & Management, vol. 11, no. 1, Art. no. 2316044, 2024, doi: 10.1080/23311975.2024.2316044.

[5] A. Zakariya and A. A. Arifin, “The Application of Digital Payments (QRIS & E-Wallet) on the Efficiency of MSME Transactions,” Journal of Finance, Economics and Business, vol. 4, no. 2, pp. 1–19, 2025, doi: 10.59827/jfeb.v4i2.326.

[6] C. P. Blocker and D. J. Flint, “Customer segments as moving targets: Integrating customer value dynamism into segment instability logic,” Industrial Marketing Management, vol. 36, no. 6, pp. 810–822, 2007, doi: 10.1016/j.indmarman.2006.05.016.

[7] S. H. Ha, S. M. Bae, and S. C. Park, “Customer's time-variant purchase behavior and corresponding marketing strategies: an online retailer's case,” Computers & Industrial Engineering, vol. 43, no. 4, pp. 801–820, 2002, doi: 10.1016/S0360-8352(02)00141-9.

[8] I. Wicaksono, “Role of Habit and Trust in Driving the Continuance of QR Code Payment Usage: A Study of QRIS Consumers in Indonesia,” Journal of Marketing Innovation (JMI), vol. 5, no. 2, 2025, doi: 10.35313/jmi.v5i2.273.

[9] A. M. Hughes, Strategic Database Marketing. Chicago, IL, USA: Probus Publishing, 1994.

[10] P. S. Fader, B. G. S. Hardie, and K. L. Lee, “RFM and CLV: Using Iso-Value Curves for Customer Base Analysis,” Journal of Marketing Research, vol. 42, no. 4, pp. 415–430, 2005, doi: 10.1509/jmkr.2005.42.4.415.

[11] M. A. Gomes and T. Meisen, “A review on customer segmentation methods for personalized customer targeting in e-commerce use cases,” Information Systems and e-Business Management, vol. 21, no. 3, pp. 527–570, 2023, doi: 10.1007/s10257-023-00640-4.

[12] H. T. Duong and T. D. Pham, “Customer segmentation for marketing and business management in electronic retailing,” International Journal of Business Innovation and Research, vol. 40, no. 5, pp. 53–71, 2026, doi: 10.1504/ijbir.2026.153810.

[13] Y. R. Rahadian et al., “Customer segmentation based on RFM, deep embedded clustering, and lifetime value analysis for loyalty strategy optimization,” Procedia Computer Science, vol. 284, pp. 1611–1618, 2026, doi: 10.1016/j.procs.2026.07.154.

[14] Y. Bai, “Research Review on Customer Life Cycle Value Prediction Based on RFM Analysis and Ensemble Learning,” in CPS Digital Library - Series of Conferences, 2026. [Online]. Available: https://seriesofconference.com/index.php/SCJ/article/download/403/401

[15] T. Herve, “Identification of the Productive Potential of Customers in a Commercial Bank Portfolio: Calculation of the Customer Lifetime Value using Statistical Learning Method,” International Journal of Science and Research (IJSR), vol. 10, no. 1, pp. 1435–1444, 2021, doi: 10.21275/sr21123105748.

[16] C. Cortes and V. Vapnik, “Support-vector networks,” Machine Learning, vol. 20, no. 3, pp. 273–297, 1995, doi: 10.1007/BF00994018.

[17] M. N. H. Mamun, “Advancements in Machine Learning for Customer Retention: A Systematic Literature Review of Predictive Models and Churn Analysis,” Journal of Sustainable Development and Policy, vol. 1, no. 1, pp. 250–284, 2025, doi: 10.63125/9b316w70.

[18] M. Imani, M. Joudaki, A. Beikmohammadi, and H. Arabnia, “Customer Churn Prediction: A Systematic Review of Recent Advances, Trends, and Challenges in Machine Learning and Deep Learning,” Machine Learning and Knowledge Extraction, vol. 7, no. 3, Art. no. 105, 2025, doi: 10.3390/make7030105.

[19] S. M. Lundberg and S.-I. Lee, “A Unified Approach to Interpreting Model Predictions,” in Advances in Neural Information Processing Systems 30, 2017, pp. 4765–4774.

[20] Y. Li and K. Yan, “Prediction of bank credit customers churn based on machine learning and interpretability analysis,” Data Science in Finance and Economics, vol. 5, no. 1, pp. 19–34, 2025, doi: 10.3934/dsfe.2025002.

[21] K. Peng, Y. Peng, and W. Li, “Research on customer churn prediction and model interpretability analysis,” PLOS ONE, vol. 18, no. 12, Art. no. e0289724, 2023, doi: 10.1371/journal.pone.0289724.

[22] L. Li, “An Experimental Study on E-commerce Customer Churn Prediction Based on the Explainable XGBoost Algorithm,” Frontiers in Computing and Intelligent Systems, vol. 17, no. 1, pp. 68–72, 2026, doi: 10.54097/44rwqq62.

[23] C. Schröer, F. Kruse, and J. M. Gómez, “A Systematic Literature Review on Applying CRISP-DM Process Model,” Procedia Computer Science, vol. 181, pp. 526–534, 2021, doi: 10.1016/j.procs.2021.01.199.

[24] F. Provost and T. Fawcett, “Data Science and its Relationship to Big Data and Data-Driven Decision Making,” Big Data, vol. 1, no. 1, pp. 51–59, 2013, doi: 10.1089/big.2013.1508.

[25] D. Donoho, “50 Years of Data Science,” Journal of Computational and Graphical Statistics, vol. 26, no. 4, pp. 745–766, 2017, doi: 10.1080/10618600.2017.1384734.

[26] G. Shmueli, “To Explain or to Predict?” Statistical Science, vol. 25, no. 3, pp. 289–310, 2010, doi: 10.1214/10-STS330.

[27] J. MacQueen, “Some methods for classification and analysis of multivariate observations,” in Proceedings of the Fifth Berkeley Symposium on Mathematical Statistics and Probability, Berkeley, CA, USA: University of California Press, 1967, vol. 1, pp. 281–297.

[28] P. J. Rousseeuw, “Silhouettes: A graphical aid to the interpretation and validation of cluster analysis,” Journal of Computational and Applied Mathematics, vol. 20, pp. 53–65, 1987, doi: 10.1016/0377-0427(87)90125-7.

[29] S. Kaufman, S. Rosset, C. Perlich, and O. Stitelman, “Leakage in data mining: Formulation, detection, and avoidance,” ACM Transactions on Knowledge Discovery from Data, vol. 6, no. 4, pp. 1–21, 2012, doi: 10.1145/2382577.2382579.

[30] C. Bergmeir and J. M. Benítez, “On the use of cross-validation for time series predictor evaluation,” Information Sciences, vol. 191, pp. 192–213, 2012, doi: 10.1016/j.ins.2011.12.028.

[31] N. V. Chawla, K. W. Bowyer, L. O. Hall, and W. P. Kegelmeyer, “SMOTE: Synthetic Minority Over-sampling Technique,” Journal of Artificial Intelligence Research, vol. 16, pp. 321–357, 2002, doi: 10.1613/jair.953.

[32] T. Chen and C. Guestrin, “XGBoost: A Scalable Tree Boosting System,” in Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, New York, NY, USA: ACM, 2016, pp. 785–794, doi: 10.1145/2939672.2939785.

[33] S. M. Lundberg et al., “From local explanations to global understanding with explainable AI for trees,” Nature Machine Intelligence, vol. 2, no. 1, pp. 56–67, 2020, doi: 10.1038/s42256-019-0138-9.

[34] L. Grinsztajn, E. Oyallon, and G. Varoquaux, “Why do tree-based models still outperform deep learning on typical tabular data?” in Advances in Neural Information Processing Systems 35, 2022, pp. 507–520, doi: 10.52202/068431-0037.

[35] T. Lan, H. A. Aldowah, and S. A. Shahani, “Machine learning analysis of telecom customer churn in California,” International Journal of Informatics and Communication Technology (IJ-ICT), vol. 15, no. 4, pp. 2000–2010, 2026, doi: 10.11591/ijict.v15i4.pp2000-2010.

[36] U. I. Hartanto, I. G. P. A. Buditjahjanto, and W. Yustanti, “Hybrid Clustering and Classification of At-Risk Customer Segments in Network Marketing,” Journal of Information Engineering and Educational Technology, vol. 9, no. 1, pp. 42–50, 2025, doi: 10.26740/jieet.v9n1.p42-50.

Additional Files

Published

2026-10-31

How to Cite

Zulham Sitorus, Hanna Willa Dhany, & Ade Iskandar. (2026). Comparison of XGBoost and SVM for Predicting User Behaviour Segment Shifts of QRIS Merchants Based on Dynamic RFM with SHAP Interpretation. JURIKOM (Jurnal Riset Komputer), 13(5), 1541–1551. https://doi.org/10.30865/jurikom.v13i5.9576