Algorithm | Dataset 1 | Dataset 2 | Dataset 3 | Dataset 4 | Average accuracy (testsets) (%) | ||||||
---|---|---|---|---|---|---|---|---|---|---|---|
Parameters | Accuracy | Parameters | Accuracy | Parameters | Accuracy | Parameters | Accuracy | ||||
Testset 1a (%) | Testset 1b (%) | Testset 2a (%) | Testset 2b (%) | Testset 3 (%) | Testset 4 (%) | ||||||
LOADDx (CTD KB) | P = 25, Q = 225 | 69.23 | 84.62 | P = 50, Q = 300 | 75 | 80 | P = 25, Q = 25 | 75 | P = 25, Q = 50 | 85.71 | 78.26 |
SCADDx (CTD KB) | P = 100, Q = 175 | 76.92 | 84.62 | P = 150, Q = 300 | 100 | 86.66 | P = 25, Q = 25 | 100 | P = 25, Q = 25 | 85.71 | \({\textbf {*88.99}}\) |
LOADDx (DisGeNet KB) | P = 275, Q = 50 | 76.92 | 92.31 | P = 100, Q = 75 | 93.75 | 93.33 | P = 25, Q = 25 | 100 | P = 25, Q = 25 | 85.71 | \({\textbf {*90.34}}\) |
SCADDx (DisGeNet KB) | P = 300, Q = 100 | 76.92 | 92.31 | P = 75, Q = 100 | 100 | 93.33 | P = 25, Q = 25 | 100 | P = 25, Q = 25 | 85.71 | \({\textbf {*91.38}}\) |
k-NN | K = 11 | 46.15 | 61.54 | K = 13 | 87.5 | 93.33 | K=1 | 75 | K=7 | 57.14 | 70.11 |
Random Forest | \(n_{a}\) = 100, \(n_{t}\) = 100 | 76.92 | 76.92 | \(n_{a}\) = 100, \(n_{t}\) = 100 | 100 | 93.33 | \(n_{a}\) = 90, \(n_{t}\) = 100 | 100 | \(n_{a}\) = 50, \(n_{t}\) = 100 | 71.43 | 86.43 |
Linear SVM | C = \(2^{-5}\) | 76.92 | 61.54 | C = \(2^{-5}\) | 100 | 93.33 | C = \(2^{-5}\) | 100 | C = \(2^{-5}\) | 71.43 | 83.87 |
SVM with RBF Kernel | \(\sigma\) = \(2^{-15}\), C = \(2^{3}\) | 76.92 | 61.54 | \(\sigma\) = \(2^{-15}\), C = \(2^{0}\) | 75 | 86.67 | \(\sigma\) = \(2^{-7}\), C = \(2^{-1}\) | 100 | \(\sigma\) = \(2^{3}\), C = \(2^{0}\) | 42.86 | 73.83 |
XGBoost (GBTree) | eta = 0.3, \(max\_depth\) = 2 nround = 30 | 76.92 | 76.92 | eta = 0.3, \(max\_depth\) = 2 nround = 20 | 100 | 93.33 | eta = 0.8, \(max\_depth\) = 2 nround = 10 | 100 | eta = 0.1, \(max\_depth\) = 2 nround = 90 | 71.43 | 86.43 |