The week's pick
Random Articles
Reseach Article
Loan Default Prediction System using Forward and Backward Propagation Techniques: A Comparative Study
| International Journal of Applied Information Systems |
| Foundation of Computer Science (FCS), NY, USA |
| Volume 13 - Number 3 |
| Year of Publication: 2026 |
| Authors: Ismail Idowu Akuji, Babajide Olanrewaju Ahmed, Taofik Abiodun Ahmed, Idris Babatunde Adeyemi, Ayodeji Jubril Alabi |
10.5120/ijais8bfa29462311
|
Ismail Idowu Akuji, Babajide Olanrewaju Ahmed, Taofik Abiodun Ahmed, Idris Babatunde Adeyemi, Ayodeji Jubril Alabi . Loan Default Prediction System using Forward and Backward Propagation Techniques: A Comparative Study. International Journal of Applied Information Systems. 13, 3 ( Jun 2026), 22-32. DOI=10.5120/ijais8bfa29462311
Abstract
In recent times, loan default has become a critical problem for financial institutions, underscoring the need for an accurate and reliable system that is capable of identifying vulnerable defaulters. This facilitates borrowers’ creditworthiness assessment, which in turn assists in mitigating potential losses. Hence, this study proposes a comparative analysis of forward and backward propagation neural networks for predicting loan defaulters. The dataset used for the models was sourced from the Kaggle repository, and it consists of 255,347 instances and 17 features, which were undersampled to 47,444, 50% Class 0 and 50% Class 1, due to the class distribution being imbalanced. The analysis of variance (ANOVA) was employed to identify the set of features that may or may not influence the target feature. K-fold cross-validation was applied to assess the robustness and generalization ability of the proposed models. Results show that both models improved substantially after cross-validation, especially in terms of accuracy and loss. The FPNN increased from 0.7110 baseline accuracy to 0.8861 mean cross-validated accuracy. At the same time, the BPNN improved from 0.6920 to 0.8858, indicating that cross-validation produced a more stable and reliable estimate of model performance. Similarly, the loss values dropped from 0.584308 to 0.314693 for FPNN and from 0.605187 to 0.314709 for BPNN, which suggests better learning and stronger generalization after validation across multiple folds. The findings of this study highlight the potential of machine learning techniques in improving loan default prediction while reducing lending risk, particularly the efficacy of cross-validation in affirming the robustness and generalization of machine learning models. Future studies can build on this research by using different datasets and integrating hyperparameter tuning to further improve models, especially in terms of precision, thereby contributing to an effective, reliable, and deployable loan default prediction model.
References
- S.T. Baidoo, H. Yusif, and E.K. Ayesu, “Improving loan repayment in Ghana”. Journal of Management Science. 1(2), 21-42, 2020.
- M. Anand, A. Velu, and P. Whig, “Prediction of loan behaviour with machine learning models for secure banking”. Journal of Computer Science and Engineering (JCSE), 3(1), 1-13, 2022.
- A.A. Egwa, B. Habeeb, A.A. Ahmad, and S.M. Bizi, “Default Prediction for Loan Lenders Using Machine Learning Algorithms”, 2022.
- Y. Lai, “Credit Default Analysis and Prediction Based on Machine Learning”. Highlights in Business, Economics and Management, 21, 782-790, 2023.
- S. Mestiri, and S.M. Hiboun, “Credit scoring using machine learning and deep Learning-Based models”, 2024.
- A. Alonso, and J.M. Carbó, “Understanding the performance of machine learning models to predict credit default: a novel approach for supervisory evaluation,” 2021.
- A. Ampountolas, T. Nyarko Nde, P. Date, and C. Constantinescu, “A machine learning approach for micro-credit scoring”. Risks, 9(3), 50, 2021.
- M.H. Khedr, N.A. Azim, and A.M. Ammar, “A New Prediction Approach for Preventing Default Customers from Applying Personal Loans Using Machine Learning”. International Journal of Computer Science and Mobile Computing, 10(12), 71-82, 2021.
- M. Madaan, A. Kumar, C. Keshri, R. Jain, and P. Nagrath, “Loan default prediction using decision trees and random forest: A comparative study”. In IOP conference series: materials science and engineering (Vol. 1022, No. 1, p. 012042). IOP Publishing, 2021.
- M.N. Alam, and M.M. Ali, “Loan default risk prediction using knowledge graph”. In 2022 14th International Conference on Knowledge and Smart Technology (KST) (pp. 34-39). IEEE, 2022.
- B. Li, “Online Loan Default Prediction Model Based on Deep Learning Neural Network”. Computational Intelligence and Neuroscience, 2022, 1–9, 2022.
- B. Chen, W. Jin, and H. Lu, “Using a genetic backpropagation neural network model for credit risk assessment in the micro, small, and medium-sized enterprises”. Heliyon, 10(14)., 2024.
- W. Sun, Y. Zhu, and Q. Hu, “Study on Credit Default Risk Prediction Model Based on BP-RF Neural Network”. In 2021 2nd International Conference on Computer Science and Management Technology (ICCSMT) (pp. 154-159). IEEE Computer Society, 2021.
Index Terms
Keywords