Table 2 Performance rates for three different ML approaches: mean (sd)
From: A machine learning approach for predicting methionine oxidation sites
Feature set | AUC | Accuracy | Sensitivity | Specificity | F-measure | MCC |
|---|---|---|---|---|---|---|
TRAINING SET | ||||||
RF | ||||||
Primary | 1.0000 (0) | 0.8957 (0.0480) | 1 (0) | 0.8807 (0.0546) | 0.7176 (0.0938) | 0.7054 (0.0920) |
Tertiary | 0.9996 (0.0003) | 0.8316 (0.0591) | 1 (0) | 0.8074 (0.0674) | 0.6096 (0.0898) | 0.5977 (0.0882) |
Whole | 1.0000 (0) | 0.8948 (0.0533) | 1 (0) | 0.8797 (0.0609) | 0.7192 (0.1053) | 0.7071 (0.1046) |
mRMR | 1.0000 (0) | 0.8932 (0.0480) | 1 (0) | 0.8777 (0.0550) | 0.7138 (0.0966) | 0.7015 (0.0960) |
SVM | ||||||
Primary | 0.9997 (0.0011) | 0.9069 (0.1990) | 0.9990 (0.0034) | 0.8939 (0.2270) | 0.8751 (0.2584) | 0.8670 (0.2747) |
Tertiary | 0.9924 (0.0090) | 0.7425 (0.1501) | 0.9865 (0.0217) | 0.7077 (0.1729) | 0.5562 (0.2335) | 0.5390 (0.2407) |
Whole | 0.9992 (0.0025) | 0.9310 (0.1542) | 0.9980 (0.0058) | 0.9210 (0.1772) | 0.8936 (0.2254) | 0.8874 (0.2370) |
mRMR | 0.9995 (0.0018) | 0.9044 (0.1766) | 0.9982 (0.0043) | 0.8907 (0.2026) | 0.8545 (0.2561) | 0.8463 (0.2690) |
NN | ||||||
Primary | 0.9482 (0.0339) | 0.7248 (0.1607) | 0.9377 (0.0416) | 0.6938 (0.1841) | 0.5195 (0.1975) | 0.4835 (0.2133) |
Tertiary | 0.9336 (0.0227) | 0.7552 (0.1040) | 0.9079 (0.0322) | 0.7334 (0.1195) | 0.5082 (0.1322) | 0.4706 (0.1378) |
Whole | 0.9616 (0.0247) | 0.8273 (0.1170) | 0.9491 (0.0327) | 0.8098 (0.1333) | 0.6292 (0.1883) | 0.6063 (0.1958) |
mRMR | 0.9533 (0.0232) | 0.7897 (0.1160) | 0.9373 (0.0314) | 0.7684 (0.1325) | 0.5696 (0.1738) | 0.5413 (0.1822) |
TESTING SET | ||||||
RF | ||||||
Primary | 0.6947 (0.0416) | 0.6207 (0.0666) | 0.6737 (0.1296) | 0.6139 (0.0883) | 0.3026 (0.0439) | 0.1936 (0.0573) |
Tertiary | 0.7614 (0.0375) | 0.6975 (0.0485) | 0.7064 (0.1029) | 0.6959 (0.0633) | 0.3638 (0.0463) | 0.2781 (0.0547) |
Whole | 0.7957 (0.0355) | 0.7458 (0.0622) | 0.6849 (0.1195) | 0.7540 (0.0813) | 0.4003 (0.0563) | 0.3205 (0.0625) |
mRMR | 0.7998 (0.0334) | 0.7468 (0.0567) | 0.6817 (0.0982) | 0.7557 (0.0721) | 0.4003 (0.0562) | 0.3190 (0.0622) |
SVM | ||||||
Primary | 0.5660 (0.0431) | 0.5604 (0.0847) | 0.5383 (0.1381) | 0.5641 (0.1112) | 0.2286 (0.0414) | 0.0688 (0.0573) |
Tertiary | 0.6480 (0.0534) | 0.6434 (0.0825) | 0.5500 (0.1329) | 0.6561 (0.1070) | 0.2741 (0.0459) | 0.1437 (0.0605) |
Whole | 0.6753 (0.0424) | 0.6441 (0.0704) | 0.6037 (0.1301) | 0.6501 (0.0954) | 0.2924 (0.0417) | 0.1744 (0.0498) |
mRMR | 0.6700 (0.0450) | 0.6348 (0.0802) | 0.5986 (0.1309) | 0.6398 (0.1047) | 0.2865 (0.0461) | 0.1641 (0.0585) |
NN | ||||||
Primary | 0.5601 (0.0479) | 0.5477 (0.0907) | 0.5465 (0.1349) | 0.5474 (0.1178) | 0.2274 (0.0411) | 0.0637 (0.0567) |
Tertiary | 0.6887 (0.0470) | 0.6662 (0.0687) | 0.5998 (0.1412) | 0.6745 (0.0907) | 0.3047 (0.0523) | 0.1907 (0.0658) |
Whole | 0.6846 (0.0469) | 0.6650 (0.0680) | 0.5793 (0.1194) | 0.6765 (0.0886) | 0.2981 (0.0453) | 0.1791 (0.0581) |
mRMR | 0.6903 (0.0486) | 0.6573 (0.0696) | 0.6101 (0.1224) | 0.6640 (0.0903) | 0.3044 (0.0474) | 0.1900 (0.0627) |