Cloud-Integrated AIoT Framework for Real-Time Credit Risk and Supply Chain Analytics: A Data generated Conceptualization based on cloud & Financial Technologies.

Description

This paper presents a cloud-integrated Artificial Intelligence of Things (AIoT) framework for real-time credit risk and supply chain analytics. The proposed system combines IoT-driven data streams with scalable cloud infrastructure to enable continuous monitoring and predictive modeling of financial and operational risks. Machine learning models, including Probability of Default (PD) and Loss Given Default (LGD), are deployed within a real-time analytics pipeline to assess dynamic risk exposure across interconnected systems. The framework supports decision intelligence through automated insights and anomaly detection. Experimental evaluation using synthetic datasets demonstrates improved predictive performance and reduced latency compared to traditional batch-processing approaches. SHAP-based explainability analysis provides transparent, regulator-compliant risk predictions achieving AUC of 0.86 and R-squared of 0.99 with gradient boosting. The results highlight the effectiveness of integrating AIoT and cloud computing for scalable, real-time business analytics in fintech and supply chain domains.

Steps to reproduce

The rapid evolution of digital ecosystems has significantly increased the volume, velocity, and variety of data generated across financial and operational systems. Traditional credit risk assessment and supply chain management systems rely heavily on static, batch-processed data, limiting their responsiveness to real-time changes. This creates inefficiencies in risk identification and decision-making, particularly in dynamic environments such as fintech and global supply chains. Recent advancements in Artificial Intelligence of Things (AIoT) and cloud computing provide an opportunity to address these limitations. AIoT enables continuous data collection through interconnected devices, while cloud platforms offer scalable processing capabilities. However, existing systems often treat financial risk analytics and supply chain monitoring as separate domains, resulting in fragmented insights and suboptimal risk detection. This paper proposes a unified cloud-integrated AIoT framework that bridges this gap by combining real-time data streams with predictive analytics. The key contributions of this study include: (1) A scalable five-layer architecture integrating IoT, cloud computing, and machine learning; (2) A real-time credit risk modeling approach using PD and LGD integrated with IoT data; (3) Integration of supply chain data into financial risk analysis for holistic risk assessment; (4) SHAP-based explainability enabling regulatory compliance and model transparency; (5) Comprehensive experimental validation demonstrating 98.4% prediction accuracy and substantial performance improvements over batch systems.

Institutions

• North South University

  Dhaka Division, Dhaka

Categories

Licence