Table 7 Comparing the performance of LOADDx and SCADDx with the performance of existing machine learning algorithms using the LOOCV approach (\(n=10\) for SCADDx and LOADDx)
Algorithm | Mean accuracy (LOOCV) | Average accuracy (datasets) (%) | |||||
|---|---|---|---|---|---|---|---|
Dataset 1a (%) | Dataset 1b (%) | Dataset 2a (%) | Dataset 2b (%) | Dataset 3 (%) | Dataset 4 (%) | ||
LOADDx (CTD KB) | 82.69 | 75 | 90.16 | 91.80 | 93.33 | 72.73 | \({\textbf {**84.29}}\) |
SCADDx (CTD KB) | 80.77 | 80.77 | 96.72 | 95.08 | 93.33 | 72.73 | \({\textbf {**86.57}}\) |
LOADDx (DisGeNet KB) | 82.69 | 78.85 | 96.72 | 95.08 | 93.33 | 72.73 | \({\textbf {**86.57}}\) |
SCADDx (DisGeNet KB) | 84.62 | 76.92 | 96.72 | 93.44 | 93.33 | 72.73 | \({\textbf {**86.29}}\) |
k-NN | 48.08 | 51.92 | 81.97 | 80.32 | 80 | 40.90 | 63.87 |
Random Forest | 80.77 | 67.31 | 90.16 | 86.88 | 86.66 | 63.64 | 79.24 |
Linear SVM | 73.08 | 75 | 90.16 | 90.16 | 73.33 | 59.09 | 76.80 |
SVM with RBF Kernel | 76.92 | 65.38 | 90.16 | 90.16 | 73.33 | 59.09 | 75.84 |
XGBoost (GBTree) | 80.77 | 69.23 | 91.80 | 88.52 | 86.66 | 54.54 | 78.59 |