Skip to main content
. 2019 Oct 28;8(11):1332. doi: 10.3390/cells8111332

Table 2.

Evaluation assessment between our proposed ensemble method and individual classifiers based on an independent test set.

Method MCC ACC SN SP AUC p-Value
4mCpred-EL 0.596 ± 0.022 0.798 ± 0.011 0.804 ± 0.012 0.792 ± 0.028 0.897 ± 0.008
RF 0.562 ± 0.010 0.780 ± 0.005 0.736 ± 0.008 0.824 ± 0.014 0.862 ± 0.002 0.179
ERT 0.552 ± 0.012 0.775 ± 0.006 0.736 ± 0.008 0.814 ± 0.008 0.862 ± 0.011 0.179
SVM 0.544 ± 0.014 0.772 ± 0.007 0.742 ± 0.014 0.801 ± 0.009 0.870 ± 0.007 0.293
GB 0.544 ± 0.011 0.772 ± 0.005 0.743 ± 0.011 0.800 ± 0.020 0.862 ± 0.004 0.179
LR 0.536 ± 0.013 0.768 ± 0.006 0.753 ± 0.005 0.783 ± 0.010 0.863 ± 0.003 0.191
AB 0.524 ± 0.009 0.762 ± 0.005 0.747 ± 0.006 0.776 ± 0.008 0.793 ± 0.017 0.000371
KNN 0.508 ± 0.017 0.747 ± 0.008 0.625 ± 0.006 0.868 ± 0.012 0.826 ± 0.005 0.010668

Note: p < 0.05 shows a statistically significant difference between 4mCpred-EL and other method that is depicted in bold. Values are expressed as mean ± standard deviation.