“…On the other hand, SMOTE improves the predictive analytics (using random forest) and offers higher gain comparing with the original data (with no augmentation) with a cost reduction up to 3% in Case II with existing similar studies. Most of the studies including, [49], [86] do not report the parameter settings of each chosen classifier which makes it a challenging task to reproduce their experiments. For instance, authors in [49] achieve a total cost of A C11,090 using RF as a binary classifier on the original data; however, they do not report parameter settings, imputation technique, whether a decision threshold is applied and its value or other evaluation metrics including precision and recall.…”