the EU, Australia, and Germany, assessed by AUC, MCC, and cost metrics (6). It lacks the qualities of hybrid adaptation. The GA–RF model outperforms these by including genetic optimisation and ensemble accuracy, hence enhancing detection robustness. This study evaluates Naïve Bayes, KNN, and Logistic Regression on imbalanced credit card datasets, yielding accuracies of 97.92%, 97.69%, and 54.86%, respectively (7). Notwithstanding efficient resampling, it is inadequate in hybrid ensemble efficiency. The GA–RF model minimises im
l detection scores across various FinTech datasets. This paper defines the concepts of Genetic Algorithms (GAs) and their applicability to optimisation issues, although it fails to include domain-specific learning models (13). The GA–RF hybrid advances this by implementing GA-driven feature optimisation, improving fraud detection accuracy and resilience