Impact of Explainable AI in Credit Scoring Analysis in SBI, NTR District: Evaluating Predictive Accuracy, Fairness and Interpretability

Author

Shaik Bushra Imtiaz, Dr. Chandu Venkateswarlu


Abstract

 The credit risk assessment process plays a vital role in financial decision-making, contributing to the approval of financial services such as loans and understanding the risks associated with them. Although machine-learning systems can sift through advanced borrower data, their mechanisms can be difficult for customers to comprehend. This transparency challenge is tackled by Explainable Artificial Intelligence (XAI), which offers explanations for model decisions. This study quantifies the relationship between XAI and perceived predictive accuracy, fairness, and interpretability in the credit scoring context, with a specific focus on the NTR District in the southern Indian state of Andhra Pradesh. The four constructs (Explainable AI, Predictive Accuracy, Fairness, and Interpretability) were measured using a structured questionnaire with a five-point Likert scale. A set of statistics containing 124 responses was provided. Descriptive statistics, Cronbach's Alpha, Pearson correlation, and simple linear regression were used. The results show positive and statistically significant relationships between XAI and predictive accuracy (r = 0.306, p < 0.001), fairness (r = 0.468, p < 0.001), and interpretability (r = 0.609, p < 0.001). All four constructs had satisfactory internal consistency (Cronbach's alpha > 0.70). The results show that explainability perceptions are correlated with accuracy, fairness, and interpretability, with higher explainability perceptions resulting in higher accuracy, fairness, and interpretability perceptions. The workbook provided is simulated; therefore, this writing is a practice or draft of the writing, and the numerical data should be changed to the result of the Academic writing in Google Form answer sheet, before academic writing is submitted. 



Keywords

Explainable Artificial Intelligence, Credit Scoring, Predictive Accuracy, Fairness, Interpretability, SHAP, LIME, NTR District.



Full Text:

Download Paper PDF


References


  1. Shreya, & Pathak, H. (2025). Explainable artificial intelligence credit risk assessment using machine learning. arXiv.
  2. Sowmiya, M. N., Jaya Sri, S., Deepshika, S., & Hanushya Devi, G. (2024). Credit risk analysis using explainable artificial intelligence. Journal of Soft Computing Paradigm, 6(3), 272–283. https://doi.org/10.36548/jscp.2024.3.004
  3. Manore, N. B., & Pathare, S. N. (2026). Explainable artificial intelligence for credit risk prediction: An interpretable machine learning approach. International Journal of Engineering Research & Technology.
  4. Nhyvija, J., Nisma, M. R., Safna, S., Karthika, M., & Hafna Fathima, M. (2026). Explainable AI for credit risk and loan approval systems. International Journal of Scientific Development and Research.
  5. Liu, D., & Feng, F. (2024). Advancing credit scoring models: Integrating explainable AI for fair and transparent financial decision-making.
  6. Choudhury, M. T. H., Zakaria, R. M., Sultana, N., & Rahman, H. (2025). Explainable artificial intelligence for credit risk assessment: Balancing transparency and predictive performance. Journal of Economics, Finance and Accounting Studies, 7(6), 14–27.
  7. Ishtiaq, W. (2025). Explainable AI models for credit card default prediction: Balancing accuracy and interpretability. International Journal of Multidisciplinary Sciences and Arts, 4(4), 233–240. https://doi.org/10.47709/ijmdsa.v4i4.6900
  8. Ogbuefi, E., Aifuwa, S. E., Olatunde-Thorpe, J., & Akokodaripon, D. (2025). Explainable AI in credit decisioning: Balancing accuracy and transparency. International Journal of Advanced Multidisciplinary Research and Studies, 5(5), 913–923.

 

 

Share your valuable work from Social Media Buttons