|
Sensitivity
|
Specificity
|
G-mean
|
---|
Datasets
|
SMOTE
|
LVQ-SMOTE
|
SMOTE
|
LVQ-SMOTE
|
SMOTE
|
LVQ-SMOTE
|
---|
Breast-w
|
76.40%
|
74.16%
|
64.21%
|
67.89%
|
70.31%
|
71.03%
|
Blood
|
95.44%
|
95.00%
|
97.38%
|
99.04%
|
96.41%
|
97.02%
|
Colon-cancer
|
80.00%
|
85.00%
|
63.64%
|
72.73%
|
71.82%
|
78.86%
|
Ionosphere
|
80.16%
|
86.51%
|
91.56%
|
92.44%
|
85.86%
|
89.48%
|
Leukemia
|
95.65%
|
100.0%
|
95.92%
|
100.0%
|
95.79%
|
100.0%
|
Pima
|
72.76%
|
71.27%
|
77.60%
|
80.20%
|
75.18%
|
75.73%
|
Satimage
|
78.75%
|
75.76%
|
68.53%
|
75.67%
|
73.64%
|
75.71%
|
Yeast
|
74.51%
|
71.72%
|
86.81%
|
90.81%
|
80.66%
|
81.27%
|
- This is the case of Logistic Tree which has shown the highest G-mean among the basic classification algorithms in Table 2.