Such dilemma inspires this study to develop an NVM-friendly bagging strategy for the random forest algorithm, in order to trade the “randomness” of the sampled data for the reduced data movements in ...