The loan approval process is a vital process within the Banking industry. A variety of factors are taken into consideration when approving a loan to an applicant. These include income, age, experience in employment, credit score, loan amount requested, education, property ownership, and previous loan history. Machine learning can be used to analyse the data and predict whether an application will be approved or denied. This research contains two parts. Part A contains a dataset with 45,000 entries of loan applications. This dataset includes a variety of information about the applicant including income, employment, credit score, loan amount, reason for loan, etc. The target column indicates whether the loan has been approved or denied. Various machine learning classification models such as Logistic Regression, Decision Tree, Random Forest, K-Nearest Neighbours (KNN), and Support Vector Machine (SVM) can be applied to the dataset. These can then be compared using accuracy, precision, recall, and F1-score to see which machine learning model works best for predicting loan approvals. Part B of the research looks at comparing the lending rates of different banks using data provided by the Reserve Bank of India (RBI). The term loan rates for different banks can be compared. This can be simplified by looking at a set loan amount and time period. We can then calculate the estimated monthly EMI using the lending rates. This will help us understand how varying lending rates can impact the monthly repayment amount. This research combines machine learning to predict loan approvals and comparing lending rates. This can help us understand the key factors that relate to loan approvals and provides a simple way of determining which bank has the cheapest lending rates.
Keywords
Loan Approval Prediction, Machine Learning, Random Forest, Credit Risk, MCLR, EMI, Data Science.
This paper studied loan approval patterns and bank lending rates using data science and machine learning. In Part A, 45,000 loan applications were studied and four classifiers were compared. Loan-to-income ratio, loan interest rate, previous-loan history and income were the most important factors, while credit score contributed little. Random Forest was the best model, with 92.74% accuracy, 89.18% precision, 76.65% recall, 82.44% F1-score and 97.41% ROC-AUC. In Part B, banks were compared on MCLR, minimum term-loan rate and estimated EMI. Standard Chartered Bank had the lowest minimum term-loan rate (1.00%) and the lowest estimated EMI (βΉ8,546.87), so it came out as the most suitable option among the banks compared. Overall, the study shows that loan approval prediction and rate comparison can be combined to help both banks and borrowers. The main limitations are the imbalanced target, the previous-loan pattern that is specific to this dataset, and the use of minimum rates and a single hypothetical loan. In future work, imbalance handling, hyper-parameter tuning and explainability methods can be added, applicant profiles can be linked with bank-specific rates, and the findings can be checked on more banks and recent data.
References
[1] J. Kruppa, A. Schwarz, G. Arminger and A. Ziegler, βConsumer credit risk: Individual probability estimates using machine learning,β Expert Systems with Applications, vol. 40, no. 13, pp. 5125β5131, 2013. [2] X. Dastile, T. Celik and M. Potsane, βStatistical and machine learning models in credit scoring: A systematic literature survey,β Applied Soft Computing, vol. 91, Art. 106263, 2020. [3] N. Uddin et al., βAn ensemble machine learning based bank loan approval predictions system with a smart application,β International Journal of Cognitive Computing in Engineering, vol. 4, pp. 327β339, 2023. [4] R. Kurniawan, βApplication of random forest algorithm on credit risk analysis,β Procedia Computer Science, vol. 234, pp. 169β176, 2024. [5] Y. Wang, Y. Zhang, Y. Lu and X. Yu, βA comparative assessment of credit risk model based on machine learning: A case study of bank loan data,β Procedia Computer Science, vol. 174, 2020. [6] D. P. Singh, A. Khare, N. Purusothaman, R. Lal, S. S. Chauhan and C. Kalaiarasan, βPredictive modeling for bank loan approval: From data to decisions,β Procedia Computer Science, 2025. [7] A. E. Khandani, A. J. Kim and A. W. Lo, βConsumer credit-risk models via machine-learning algorithms and risk management,β Journal of Banking & Finance, vol. 34, no. 11, pp. 2767β2787, 2010. [8] S. Lessmann, B. Baesens, H.-V. Seow and L. C. Thomas, βBenchmarking state-of-the-art classification algorithms for credit scoring: An update of research,β European Journal of Operational Research, vol. 247, no. 1, pp. 124β136, 2015. [9] L. Breiman, βRandom forests,β Machine Learning, vol. 45, no. 1, pp. 5β32, 2001. [10] F. Pedregosa et al., βScikit-learn: Machine learning in Python,β Journal of Machine Learning Research, vol. 12, pp. 2825β2830, 2011. [11] N. V. Chawla, K. W. Bowyer, L. O. Hall and W. P. Kegelmeyer, βSMOTE: Synthetic minority over-sampling technique,β Journal of Artificial Intelligence Research, vol. 16, pp. 321β357, 2002. [12] D. J. Hand and W. E. Henley, βStatistical classification methods in consumer credit scoring: A review,β Journal of the Royal Statistical Society: Series A, vol. 160, no. 3, pp. 523β541, 1997. [13] B. Baesens, T. Van Gestel, S. Viaene, M. Stepanova, J. Suykens and J. Vanthienen, βBenchmarking state-of-the-art classification algorithms for credit scoring,β Journal of the Operational Research Society, vol. 54, no. 6, pp. 627β635, 2003. [14] Reserve Bank of India, βMarginal Cost of Funds based Lending Rate (MCLR) β Guidelines,β Mumbai, India, 2016. [15] W. McKinney, βData structures for statistical computing in Python,β in Proc. 9th Python in Science Conf., 2010, pp. 56β61.
π How to Cite This Paper
Ritika Sharma, Dr. Preeti Tiwari (2026). Predictive Analytics for Loan Approval and Comparative Bank Interest Rate Evaluation. International Journal of Computer Science Engineering Techniques, 13(6), 477β487. ISSN: 2455-135X. DOI: https://doi.org/10.5281/zenodo.24369501