Research author
Marcellin Atemkeng
2 AI research papers in the One9Founders library, with summaries and links to original sources.
Papers by Marcellin Atemkeng
Balancing Performance and Fairness in Explainable AI for Anomaly Detection in Distributed Power Plants Monitoring
Corneille Niyonkuru, Marcellin Atemkeng, Gabin Maxime Nguegnang, et al.
This paper presents a machine learning system to detect equipment problems in diesel generators used by telecom operators in Cameroon. The system combines multiple AI models with special techniques to handle imbalanced data, explain its decisions through SHAP analysis, and ensure fair predictions across different regions—achieving 99% accuracy while maintaining minimal bias. The researchers also discuss how to deploy this system for real-time monitoring in practice.
An Optimised Greedy-Weighted Ensemble Framework for Financial Loan Default Prediction
Ezekiel Nii Noye Nortey, Jones Asante-Koranteng, Marcellin Atemkeng, et al.
This paper presents an Optimised Greedy-Weighted Ensemble framework for predicting loan defaults in credit risk management. The system combines multiple machine learning models with optimized weights and uses a neural network to learn how to best blend their predictions together. Testing on real lending data shows this approach achieves strong performance (80% AUC) and identifies key factors like income and debt ratios as important default predictors.