Skip to main content

Advertisement

Table 1 Mean and standard deviation (SD) values of the balanced accuracy and Matthews Correlation Coefficient (MCC) for all five learning algorithms in 20 bootstrap test samples. The best values in each fold category are underlined with the overall best in bold

From: Prediction of novel mouse TLR9 agonists using a random forest approach

AlgorithmCross-validationMean balanced accuracySD balanced accuracyMean MCCSD MCC
RF5-fold77.8%0.080.570.15
GBM5-fold76.8%0.070.550.12
SDA5-fold74.6%0.080.500.14
SVM5-fold77.1%0.080.550.16
NN5-fold74.1%0.070.500.13
RF10-fold78.9%0.060.600.11
GBM10-fold77.7%0.050.570.10
SDA10-fold75.8%0.060.530.11
SVM10-fold78.4%0.050.580.11
NN10-fold72.9%0.050.480.10
RF15-fold77.0%0.060.550.11
GBM15-fold76.9%0.060.550.11
SDA15-fold73.5%0.060.490.11
SVM15-fold76.3%0.050.530.11
NN15-fold72.6%0.070.470.15
RF20-fold79.7%0.050.610.09
GBM20-fold78.5%0.070.580.12
SDA20-fold76.1%0.080.540.14
SVM20-fold75.4%0.050.520.09
NN20-fold74.9%0.070.520.13