Table 5 Paired rank comparison of algorithms in Acc, AUC, sensitivity, specificity, and precision of predictive model built on clinic risk indicators to PVST detected by algorithms

RFA-PVST

1.7308

1.1923

1.1923

1.5769

1.0000

mRMR

*

−0.5385

−0.5385

− 0.1538

− 0.7308

SVM-RFE

0

0.3846

−0.1923

Relief

0.3846

−0.1923

S-weight

−0.5769

LLEScore

AUC

RFA-PVST

mRMR

SVM-RFE

Relief

S-weight

LLEScore

RFA-PVST

1.7308

2.4231

1.7308

1.8077

1.7692

mRMR

0.6923

0

0.0769

0.0385

SVM-RFE

*

−0.6923

−0.6154

− 0.6538

Relief

0.0769

0.0385

S-weight

−0.0385

LLEScore

Sensitivity

RFA-PVST

mRMR

SVM-RFE

Relief

S-weight

LLEScore

RFA-PVST

−0.5769

−2.2692

−2.2692

−1.8077

0.4615

mRMR

−1.6923

−1.6923

−1.2308

1.0385

SVM-RFE

*

0

0.4615

2.7308

Relief

*

0.4615

2.7308

S-weight

*

2.2692

LLEScore

*

*

*

Specificity

RFA-PVST

mRMR

SVM-RFE

Relief

S-weight

LLEScore

RFA-PVST

1.3077

2.9615

2.9615

2.6923

0.2308

mRMR

1.6538

1.6538

1.3846

−1.0769

SVM-RFE

*

0

−0.2692

−2.7308

Relief

*

−0.2692

−2.7308

S-weight

*

−2.4615

LLEScore

*

*

*

Precision

RFA-PVST

mRMR

SVM-RFE

Relief

S-weight

LLEScore

RFA-PVST

1.8462

1.8846

1.8846

2.3077

1.0769

mRMR

*

0.0385

0.0385

0.4615

−0.7692

SVM-RFE

*

0

0.4231

−0.8077

Relief

*

0.4231

−0.8077

S-weight

*

−1.2308

LLEScore