Fast and Efficient Real-Time Banking Fraud Detection Using Stream-Based Algorithms

Author(s)	Dr. Neha Patwari, Dr. Neeta Patil, Ms. Mary Margarat Neela, Ms. Jisha Tinsu
Country	India
Abstract	The efficacy of traditional batch fraud detection methods is diminishing due to the rise in digital financial transactions. To address contemporary banking fraud, this research explores real-time stream processing. It investigates memory-efficient and low-latency techniques such as One-Class SVM, Incremental Local Outlier Factor, and Isolation Forest. Additionally, the study analyzes real-time data processing methods including sampling, filtering, and approximation counting techniques like Bloom Filters and Count-Min Sketch. It discusses how existing frameworks utilizing Apache Kafka, Apache Flink, FastAPI, and Redis can be integrated to combine these various techniques as components. Models such as Random Forest and LightGBM, which effectively balance accuracy and performance during transaction processing, are found to be the most effective in detecting fraud, based on experiments conducted on a publicly available credit card dataset.
Keywords	Real-Time Fraud Detection, Stream Processing, Anomaly Detection, Apache Flink, Low Latency, Count-Min Sketch, Bloom Filter.
Field	Engineering
Published In	Volume 7, Issue 6, November-December 2025
Published On	2025-11-26
DOI	https://doi.org/10.36948/ijfmr.2025.v07i06.61545
Short DOI	https://doi.org/hbdrg7

About IJFMR Fees & Payment Current Issue Publication Archive	Submit Research Paper Track Submission Status Publication Guidelines Publication Ethics Peer Review & Plagiarism	Join as a Reviewer Editors & Reviewers Reviewer Referral Program Get Reviewer Membership Certi.	Website/Journal Policies Usage Policy Content Policies Privacy Policy

Contact Us		+91-9687-828-838	editor@ijfmr.com

International Journal For Multidisciplinary Research