Claim Missing Document
Check
Articles

Found 2 Documents
Search

Implementing Adaptive Credit Scoring Through Xgboost and Deep Learning in Open Finance for Microcredit Reza Ade Putra; Silviana Pebruary
Start-up and Financial Technology Vol. 1 No. 2 (2025)
Publisher : Generate Digital Publishing

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.70764/gdpu-sft.2025.1(2)-06

Abstract

Objective: This study aims to develop and test the effectiveness of an AI-based adaptive credit scoring system that integrates Open Finance principles and the use of alternative data in assessing microcredit eligibility in emerging markets.Research Design & Methods: This study uses a mixed-methods approach with AI model experiments to test machine learning and deep learning-based adaptive credit scoring systems and thematic analysis to understand implementation challenges and regulatory preparedness.Findings: Hypothetical results show that AI models (XGBoost, Deep Learning) significantly outperform traditional models (logistic regression) in microcredit eligibility assessment, with an AUC improvement of 15-20% (e.g., AUC 0.88 for XGBoost vs. 0.72 for logistic regression). Similar improvements are also seen in Precision, Recall, and F1-score metrics. Adaptivity testing shows a 2-5% increase in model accuracy after each retraining cycle, with the model's ability to dynamically adjust credit scores in real-time to changes in user behaviour. Interpretation of the results through XAI identified consistent utility bill payment patterns, digital wallet transaction frequency, and mobile phone number usage duration as the most influential predictive factors.Implications & Recommendations: Integration of AI and Open Finance with alternative data improves the efficiency and inclusiveness of credit scoring, so it is recommended that industry players and regulators develop data infrastructure, apply Responsible AI principles, and formulate comprehensive and collaborative Open Finance regulations.Contribution & Value Added: This research contributes to the FinTech and AI literature by providing technical and policy guidance to promote financial inclusion by developing adaptive assessment machine and responsible AI oversight in emerging markets.
Academic Risk Analysis of University Students Using Random Forest and SHAP Nouval Khairi; Reza Ade Putra
Journal of Computer Networks, Architecture and High Performance Computing Vol. 8 No. 4 (2026): Research Paper October 2026
Publisher : Information Technology and Science (ITScience)

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.47709/cnahpc.v8i4.9625

Abstract

 Academic risk classification can help universities identify students who require closer academic monitoring, but imbalanced classes and limited model transparency can reduce practical usefulness. Objective: This study develops a Random Forest classifier for academic-risk categories and explains model decisions using SHAP. Methods: The study used academic data from 484 Computer Science students at Universitas Islam Negeri Sumatera Utara for the 2025/2026 odd-to-even academic period. The target was formed using rule-based indicators, while the final predictors were Jumlah_MK (number of courses), Praktikum (practicum courses), and MK_Berat (heavy courses). Data were split stratified into 387 training and 97 testing observations. SMOTE was applied only to training folds during stratified five-fold cross-validation and again to the complete training set for the final model. Results: Cross-validation produced mean Accuracy 91.72%, Precision 46.05%, Recall 81.33%, F1-Score 58.21%, and ROC-AUC 90.46%. On the original testing set, the model achieved Accuracy 97.94%, Precision 85.71%, Recall 85.71%, F1-Score 85.71%, and ROC-AUC 98.65%, with TN=89, FP=1, FN=1, and TP=6. Feature Importance and Mean Absolute SHAP both ranked Jumlah_MK as the most influential predictor. Conclusion: Random Forest with training-only SMOTE provided strong testing performance and improved minority-class detection, while SHAP added an interpretable layer. The results should be interpreted cautiously because the high-risk class was small and the study used a single academic period