Table 1 Comparison of various predictors based on accuracy and F1-score fitness function with feature selection in training set.
From: Development of machine learning-based predictors for early diagnosis of hepatocellular carcinoma
Predictors | NO.Opt | NO.HCC | NO.CwoHCC | ACC | F1-score | AUC | 95% CI |
|---|---|---|---|---|---|---|---|
mRMR + KNN | 11 | 988/988 | 332/332 | 1 | 1 | 1 | 1–1 |
mRMR + SVM | 11 | 988/988 | 332/332 | 1 | 1 | 1 | 1–1 |
mRMR + LR | 15 | 988/988 | 332/332 | 1 | 1 | 1 | 1–1 |
mRMR + XGBoost | 26 | 988/988 | 332/332 | 1 | 1 | 1 | 1–1 |
mRMR + LMT | 26 | 988/988 | 332/332 | 1 | 1 | 1 | 1–1 |
mRMR + AdaboostM1 | 60 | 987/988 | 330/332 | 0.9977 | 0.9985 | 0.9965 | 0.9922–1 |
mRMR + J48 | 66 | 987/988 | 329/332 | 0.997 | 0.998 | 0.995 | 0.9898–1 |
mRMR + NB | 24 | 980/988 | 330/332 | 0.9924 | 0.9949 | 0.9929 | 0.9879–0.998 |
MRMD + KNN | 28 | 988/988 | 332/332 | 1 | 1 | 1 | 1–1 |
MRMD + SVM | 28 | 988/988 | 332/332 | 1 | 1 | 1 | 1–1 |
MRMD + LR | 30 | 988/988 | 332/332 | 1 | 1 | 1 | 1–1 |
MRMD + LMT | 74 | 988/988 | 332/332 | 1 | 1 | 1 | 1–1 |
MRMD + J48 | 59 | 987/988 | 329/332 | 0.997 | 0.998 | 0.995 | 0.9898–1 |
MRMD + AdaboostM1 | 160 | 985/988 | 330/332 | 0.9962 | 0.9975 | 0.9955 | 0.991–1 |
MRMD + XGBoost | 96 | 982/988 | 326/332 | 0.9909 | 0.9939 | 0.9879 | 0.9804–0.9955 |
MRMD + NB | 28 | 963/988 | 328/332 | 0.978 | 0.9852 | 0.9813 | 0.9737–0.989 |
