Skip to main content

Table 3 Models comparison. T-test p-value from bootstrap results on the testing sets

From: A machine learning approach for predicting methionine oxidation sites

Feature set RF-SVM RF-NN SVM-NN
AUC
Primary 1.337807e-53 1.656090e-52 3.629288e-01
Tertiary 7.466593e-08 7.749183e-10 3.076722e-01
Whole 1.620777e-11 1.207725e-10 6.687422e-01
mRMR 5.736385e-04 1.122952e-05 3.027066e-01
Accuracy
Primary 7.466593e-08 7.749183e-10 3.076722e-01
Tertiary 1.620777e-11 1.207725e-10 6.687422e-01
Whole 5.736385e-04 1.122952e-05 3.027066e-01
mRMR 1.110837e-35 7.810002e-38 5.302538e-01
Sensitivity
Primary 4.838807e-26 9.923600e-27 8.419212e-01
Tertiary 7.067182e-08 2.630463e-04 3.507737e-02
Whole 3.771161e-17 6.241079e-09 1.096249e-02
mRMR 1.650447e-03 5.410619e-02 1.924156e-01
Specificity
Primary 7.035627e-39 8.036713e-20 3.721847e-07
Tertiary 1.059365e-30 6.066923e-15 1.807435e-05
Whole 1.072350e-21 9.069624e-16 3.319756e-02
mRMR 7.569818e-06 2.475176e-09 1.699726e-01
F-measure
Primary 1.900064e-14 8.911586e-10 4.341598e-02
Tertiary 1.385330e-43 3.632520e-39 5.440616e-01
Whole 8.253875e-35 1.802488e-31 3.612268e-01
mRMR 5.984561e-23 4.366711e-19 3.520361e-02
MCC
Primary 9.137039e-07 8.985178e-06 5.212807e-01
Tertiary 1.701737e-16 1.821765e-13 8.146449e-02
Whole 6.201029e-44 2.659090e-33 2.914253e-03
mRMR 4.996287e-36 3.033082e-28 7.392408e-03