High-Performance Machine Learning for Large-Scale Data Classification considering Class Imbalance

Table 3

The comparisons of the classification accuracy using Borderline-SMOTE.

Imbalance class ratio	Balanced dataset accuracy (%)	Imbalanced dataset accuracy (%)
5000 : 1000	98.09	73.15
5000 : 500	95.19	68.32
5000 : 50	80.53	56.20
5000 : 10	67.45	53.13