In which scenario might you prefer using RMS over MAE? a. When the data contains many outliers b. When all errors need to be treated equally, regardless of their size c. When large errors are particularly undesirable and should be heavily penalized d. When the distribution of error needs to be skewed