Table 1 The prediction accuracy obtained by using different machine learning algorithms
From: QSAR analysis of VEGFR-2 inhibitors based on machine learning, Topomer CoMFA and molecule docking
Data set | Algorithm | SP | SN | ACC |
|---|---|---|---|---|
Training Set | Adaboost | 87.7 | 84.0 | 85.7 |
Bagging | 75.7 | 81.1 | 78.6 | |
RandomForest | 88.5 | 86.2 | 87.3 | |
RandomTree | 77.0 | 82.5 | 81.5 | |
C4.5 | 77.0 | 82.5 | 79.9 | |
ADTree | 75.7 | 69.1 | 72.2 | |
KNN | 85.2 | 88 | 87.3 | |
Bayes Net | 80.7 | 69.5 | 74.7 | |
SVM | 63.0 | 75.6 | 69.7 | |
ANN | 79.8 | 86.5 | 83.4 | |
Test Set | Adaboost | 86.1 | 81.7 | 83.7 |
Bagging | 75.5 | 76.1 | 75.8 | |
RandomForest | 85.4 | 79.4 | 82.2 | |
RandomTree | 78.1 | 73.9 | 75.8 | |
C4.5 | 67.5 | 78.9 | 73.7 | |
ADTree | 70.9 | 70.6 | 70.7 | |
KNN | 82.1 | 83.9 | 84.3 | |
Bayes Net | 78.1 | 58.7 | 67.1 | |
SVM | 60.3 | 76.7 | 69.2 | |
ANN | 76.8 | 80.6 | 78.9 |