Abstract :

Financial Ecosystems Driven By Digital Transactions And Internet Of Things (IoT) Environments Are Increasingly Vulnerable To Fraudulent Activities, Making Efficient Fraud Detection A Major Concern. The High Volume Of Transaction Data, Combined With Class Imbalance And Continuously Evolving Fraud Patterns, Makes Accurate Detection Challenging, While Traditional Rule-based And Manual Monitoring Systems Fail To Adapt To Such Dynamic Scenarios. To Address These Limitations, The Proposed System Introduces A Blockchain-enabled Incremental Learning Framework For Fraud Detection. The Framework Evaluates Multiple Machine Learning (ML) Models, Including Passive Aggressive Classifier (PAC), Stochastic Gradient Descent (SGD) Classifier, Perceptron, Naïve Bayes (NB), And Light Gradient Boosting Machine (LGBM), For Binary Classification Using The Target Column Is Fraud, Where Transactions Are Labeled As Normal Or Fraudulent. The System Adopts A Two-phase Learning Strategy Consisting Of Initial Training And Incremental Updates, Where Models Are First Trained On Historical Data And Then Continuously Updated With New Incoming Data Without Retraining From Scratch, Enabling Adaptability To Evolving Fraud Patterns. To Address Class Imbalance, Synthetic Minority Oversampling Technique (SMOTE) Is Applied, Improving Model Performance On Minority Fraudulent Cases. Among All Models, The SGD Classifier Is Selected As The Final Model Due To Its Computational Efficiency And Strong Suitability For Incremental Learning. Furthermore, Blockchain Technology Is Integrated To Securely Store And Manage Model Parameters, Ensuring Data Integrity, Transparency, And Resistance To Tampering. The System Is Implemented Using The Flask Web Framework, Providing An Interactive Interface For Realtime Fraud Prediction. The Proposed Approach Enhances Detection Accuracy, Reduces Computational Overhead, And Offers A Scalable, Secure, And Reliable Solution For Modern Financial And IoT-based Applications.

Published:

24-4-1-2026

Issue:

Vol. 26 No. 4-1 (2026)

Page Nos:

773-780

Section:

Articles

License:

This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

How to Cite