Current research work aims to explore three different ensemble optimization techniques namely Base Ensemble, Dynamic Ensemble Selection (DES), and Greedy Optimization. This analysis is carried out to determine the trade-off among predictive accuracy and the computational requirement in Ensemble modeling. This work also seeks to investigate optimization paradigms –Dynamic Ensemble Selection Performance (DESP), K-Nearest Oracles Eliminate (KNORA-E) and K-Nearest Oracles Union (KNORA-U) in the context of DES and growing and pruning strategies based on Greedy Optimization. Considered ensemble optimization techniques have been experimentally implemented on the credit card dataset and the results reveal that Greedy Optimization with growing and pruning strategies outperforms the other ones by achieving the highest accuracy of 0.72. Thus, results advocate the significance of greedy optimization and hence can be implemented in various applications including security-critical applications such as identification of threat mitigations that requires high accuracy and low computational power.