International Conference on Data Management, Analytics and Innovation (ICDMAI) · 2026 · In Press
Minority Class Over-Sampling Using Random Forest Proximity Matrix and Genetic Algorithms
Proposed a Random Forest Proximity-based Oversampling (RF-POS) technique that outperformed SMOTE across 18 benchmark datasets (MCC 91.61% vs. 90.93%), with Genetic Algorithms for synthetic-sample selection improving MCC by a further 2.09%.
Accepted 2025, in press. Placed 3rd in the Best Paper category at ICDMAI 2026.