Ponnuviji, N. P.  <br/><br/>
Morphic computing with machine learning for enhanced fraud detection in financial applications 

 - Vol.15(2), Oct 
 - Mumbai  ICT Academy  2024 
 - 3545-3550p. 
<br/><br/> As financial fraud becomes increasingly complex, traditional detection<br/>methods struggle to keep pace, resulting in substantial financial losses<br/>globally. Morphic computing—a paradigm that emphasizes adaptable,<br/>context-aware processing—offers promising advancements for fraud<br/>detection in dynamic environments. Integrating morphic computing<br/>with machine learning models creates a responsive framework capable<br/>of discerning subtle and evolving fraud patterns. The proposed system<br/>utilizes a Convolutional Neural Network (CNN) enhanced with<br/>Morphic Layering, where layers adaptively morph in response to new<br/>data patterns. The dataset, sourced from real-time financial<br/>transactions, consists of 500,000 records, including 2,000 flagged<br/>fraudulent cases. The system was tested on a simulated environment<br/>over a six-month period, yielding an accuracy of 98.5% in fraud<br/>detection and reducing false positives by 40% compared to traditional<br/>machine learning models. Latency for real-time detection was<br/>minimized to 200 milliseconds, proving feasible for immediate<br/>application in transaction monitoring systems. By offering a flexible<br/>structure, this method surpasses existing approaches, as it continuously<br/>evolves to detect emerging fraud patterns, thus enhancing financial<br/>security. 
<br/><br/><label>Subjects--Topical Terms: </label><br/>Computer Engineering