header

Articles

Vol. 1 No. 1 (2026): July 2026

Real Time Fraud Detection in Indonesian Fintech Transactions Using Apache Kafka and Streaming Machine Learning with Concept Drift Adaptation: A Systematic Literature Review

Submitted
July 3, 2026
Published
2026-07-04

Abstract

The proliferation of digital financial transactions in Indonesia has created an urgent need for robust, real-time fraud detection systems capable of adapting to continuously evolving fraud patterns. Conventional batch-processing fraud detection models are inadequate for dynamic streaming environments, particularly in the context of Indonesian fintech, where local transaction behaviors exhibit unique temporal and behavioral patterns not well represented in global datasets. This Systematic Literature Review (SLR) investigates the intersection of three critical domains: real-time data streaming architectures (specifically Apache Kafka), streaming machine learning algorithms, and concept drift adaptation mechanisms, applied to fraud detection in Indonesian fintech ecosystems. Following PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines, we conducted a comprehensive search across six academic databases—Scopus, IEEE Xplore, Web of Science, ScienceDirect, ACM Digital Library, and MDPI—covering publications from 2020 to 2024. From an initial pool of 847 articles, 42 studies met the inclusion criteria after rigorous screening. Our analysis reveals that while Apache Kafka has been increasingly adopted as a high-throughput event streaming backbone for fraud detection, its integration with adaptive streaming machine learning models that explicitly address concept drift remains critically underexplored, especially for emerging market contexts like Indonesia. Findings indicate that Adaptive Random Forest (ARF) and Hoeffding Adaptive Trees (HAT), combined with drift detectors such as ADWIN and EDDM, demonstrate superior performance in non-stationary financial streaming environments. We identify a significant research gap: the absence of locally calibrated, concept-drift-aware streaming ML frameworks for Indonesian fintech transaction patterns. This review contributes a structured research taxonomy, identifies key challenges including data imbalance, regulatory constraints (OJK/BI-FAST standards), and latency requirements, and proposes a conceptual architecture for future empirical work.

References

  1. B. Indonesia, “Laporan Sistem Pembayaran dan Pengelolaan Uang Rupiah 2023,” Bank Indonesia, Jakarta, 2024. [Online]. Available: https://www.bi.go.id
  2. Google, Temasek, and B. & Company, “e-Conomy SEA 2024: Advancing Towards a Digital Decade,” 2024. [Online]. Available: https://economysea.withgoogle.com
  3. I. A.-S. C. (IASC) and O. J. K. (OJK), “AI-Based Fraud Report: Deepfake and Digital Fraud Trends 2024–2025,” OJK, Jakarta, 2025.
  4. S. Malviya, “Limitations of batch fraud detection techniques in dynamic financial networks,” Int. J. Multidiscip. Res., vol. 11, no. 1, 2025, [Online]. Available: https://www.ijfmr.com/papers/2025/January/IJFMR0112345.pdf
  5. K. S. Hebbar, H. Parnerkar, and S. Gondi, “AI-driven real-time fraud detection using Kafka Streams in FinTech,” Int. J. Appl. Math., vol. 38, no. 6s, pp. 770–782, 2025.
  6. H. O. Bello, A. B. Ige, and M. N. Ameyaw, “Adaptive machine learning models: Concepts for real-time financial fraud prevention in dynamic environments,” World J. Adv. Eng. Technol. Sci., vol. 12, no. 02, pp. 21–34, 2024.
  7. K. Waehner, “Fraud detection with Apache Kafka, ksqlDB and Apache Flink,” 2022, Kai Waehner Blog. [Online]. Available: https://www.kai-waehner.de/blog/2022/10/25/fraud-detection-with-apache-kafka-ksql-and-apache-flink/
  8. V. Yelleti, “ROSFD: Robust online streaming fraud detection with resilience to concept drift in data streams,” arXiv, 2025. [Online]. Available: https://arxiv.org/abs/2504.10229
  9. M. K. H. Chy, “Proactive fraud defense: Machine learning’s evolving role in protecting against online fraud,” arXiv, 2024. [Online]. Available: https://arxiv.org/abs/2410.20281
  10. N. Arora and S. Arora, “Real-time fraud detection using machine learning and stream processing,” Int. J. Adv. Res. Sci. Commun. Technol., vol. 2, no. 1, pp. 142–158, 2022.
  11. J. Montiel et al., “River: Machine learning for streaming data in Python,” J. Mach. Learn. Res., vol. 22, no. 110, pp. 1–8, 2021.
  12. J. Montiel, J. Read, A. Bifet, and T. Abdessalem, “Scikit-Multiflow: A multi-output streaming framework,” J. Mach. Learn. Res., vol. 19, no. 72, pp. 1–5, 2018.
  13. A. L. Suárez-Cetrulo, D. Quintana, and A. Cervantes, “A survey on machine learning for recurring concept drifting data streams,” Expert Syst. Appl., vol. 213, p. 118934, 2023, doi: 10.1016/j.eswa.2022.118934.
  14. F. Bayram, B. S. Ahmed, and A. Kassler, “From concept drift to model degradation: An overview on performance-aware drift detectors,” Knowledge-Based Syst., vol. 245, p. 108632, 2022, doi: 10.1016/j.knosys.2022.108632.
  15. A. Bifet and R. Gavalda, “Learning from time-changing data with adaptive windowing BT - Proceedings of the 2007 SIAM International Conference on Data Mining,” SIAM, 2007, pp. 443–448. doi: 10.1137/1.9781611972771.42.
  16. M. Baena-García, J. del Campo-Ávila, R. Fidalgo, A. Bifet, R. Gavalda, and R. Morales-Bueno, “Early drift detection method BT - Proceedings of the 4th ECML PKDD International Workshop on Knowledge Discovery from Data Streams,” 2006, pp. 77–86.
  17. M. J. Page et al., “The PRISMA 2020 statement: An updated guideline for reporting systematic reviews,” BMJ, vol. 372, p. n71, 2021, doi: 10.1136/bmj.n71.
  18. S. Saket, V. Chandela, and M. D. Kalim, “Real-time event joining in practice with Kafka and Flink,” arXiv, 2024. [Online]. Available: https://arxiv.org/abs/2410.15533
  19. H. M. Gomes et al., “Adaptive random forests for evolving data stream classification,” Mach. Learn., vol. 106, no. 9–10, pp. 1469–1495, 2017, doi: 10.1007/s10994-017-5642-8.

Downloads

Download data is not yet available.