A Comparison of Two Approaches to Data Mining from Imbalanced Data