Methods CC on Training Dataset CC on Independent test
Gumbel distribution [16] 0.34 0.37
Vihinen’s methods [16] 0.31 0.33
Karplus and Schulz ’s methods [16] 0.30 0.33
NS [26] 0.34 -
KS [26] 0.38 -
PS [26] 0.41 -
2-stage SVR [12] 0.53 0.55
This worka 0.60 0.41
This work (combination)b 0.57 0.58
athe results are from the method with the best CC value on training dataset in Table 3, e.g. the Random Forest with 30 attributes bthe results are from the Random Forest with 30 attributes in Table 4. The CC values on training dataset is the average CC values of the five inner folds, and the CC value of the independent test is the related outer folds
Table 5: The comparison with other works.