Research on the Application of Big Data Cloud Scoring in Inclusive Financial Risk Identification: A Credit Assessment Framework for Thin-File and No-File Customer Segments
DOI:
https://doi.org/10.71222/dasgab89Keywords:
inclusive finance, credit scoring, big data analyticsAbstract
Inclusive financial systems face persistent challenges in extending credit to thin-file and no-file customer segments-individuals with insufficient or absent formal credit histories. Traditional scoring models fail these populations due to data scarcity, leading to systemic exclusion and underbanking. This study introduces a novel credit assessment framework grounded in big data cloud scoring, which synthesizes heterogeneous alternative data streams-including mobile transaction footprints, utility payment behaviors, social network metadata, and geospatial mobility patterns-within a federated cloud architecture. The framework employs a multi-stage risk identification pipeline: real-time data ingestion via edge-compatible APIs; privacy-preserving feature engineering using differential privacy and homomorphic encryption; adaptive ensemble modeling combining graph neural networks for relational inference and temporal convolutional networks for behavioral sequence analysis; and dynamic threshold calibration informed by local economic volatility indices. Validated across 126,480 anonymized applicants from three emerging-market regions over an 18-month period, the framework achieved 89.3% precision in default prediction at 75% recall, reducing false rejection rates by 41.7% compared to baseline bureau-based models. Notably, it demonstrated robust performance across zero-credit-history cohorts (AUC = 0.842) and maintained calibration stability despite regional inflation fluctuations exceeding 12% annually. The architecture supports regulatory compliance through auditable feature provenance logs and on-demand model explainability dashboards. These findings establish big data cloud scoring not as a substitute but as a structural complement to conventional credit infrastructure-enabling scalable, equitable, and resilient risk identification for financially underserved populations.References
1. D. Shah and C. Bhatt, "Responsible AI in Fintech: Evaluating Fairness and Bias in Rural Credit Scoring," IEEE Access, vol. 12, pp. 1–15, 2026.
2. Z. Xie and L. Zhang, "FedAttn-Credit: attention-augmented federated learning with adaptive differential privacy for rural inclusive finance credit assessment," Sci. Rep., vol. 16, p. 45823, 2026.
3. S. Purwanto, N. E. R. Wuryandari, D. H. Perkasa, and A. Kurniawan, "Big Data Credit Scoring and MSME Credit Access: Evidence from Fintech Lending in Indonesia," IEEE Trans. Eng. Manage., vol. 78, no. 1, pp. 21–35, 2026.
4. I. Olabintan, "FairLend-Africa: An Explainable Machine Learning Framework for Alternative Credit Scoring Using Behavioral Financial Data in Financially Excluded African Communities," SSRN Electron. J., doi: 10.2139/ssrn.6837665, Mar. 2026.
5. R. Gentyala, "A Hybrid Machine Learning Approach for Credit Scoring: Integrating Traditional Financial History with Mobile Phone Behavioral Metrics," Int. J. Artif. Intell. Mach. Learn. Res. Dev., vol. 3, no. 1, pp. 13–40, 2022.
6. J. A. Kumar and S. R. Babu, "Enhancing Credit Scoring with Alternative Data and Machine Learning for Financial Inclusion," IEEE Trans. Comput. Soc. Syst., vol. 11, no. 4, pp. 1–14, 2026.
7. C. K. Gali, "Algorithmic Bias in AI-Based Credit Scoring Systems: Financial Inclusion, Risk Modeling, and Ethical Constraints," IEEE Trans. Technol. Soc., vol. 7, no. 2, pp. 1–12, Mar. 2026.
8. D. Shukla and S. Gupta, "Exploring Alternative Datasets for Credit Scoring of Thin-File Consumers: A Comprehensive Review," J. Comput. Sci. Eng. Res., vol. 9, no. 3, pp. 1–22, 2025.
9. R. Noreen, M. Sajid, and F. Amjad, "Algorithmic Bias and Fairness in AI Credit Scoring: Evidence, Mechanisms, and Governance Responses," J. Manag. Sci. Res. Rev., vol. 5, no. 2, pp. 1546–1558, 2026.
10. A. A. Sabeena et al., "Using Alternative Data and Machine Learning for Predictive Credit Scoring to Promote Financial Inclusion in the US," in Proc. IEEE Int. Conf. Big Data (BigData), 2026, pp. 1–12.
11. O. Israel, "The Integration of Big Data Analytics and Machine Learning in Credit Scoring for Expanding Financial Inclusion," IEEE Trans. Knowl. Data Eng., vol. 38, no. 5, pp. 1–16, 2026.
12. F. Pramesti, H. M. Anhar, and I. Mawaddah, "Algorithmic Bias in Artificial Intelligence--Based Credit Scoring and Its Implications for Financial Inclusion," Capitalis: J. Econ. Stab., Bank., Invest., vol. 1, no. 2, pp. 12–22, 2026.
