Table 4 Performance metrics in terms of \(\text{F}_1\) and AUC by the different classifiers using radiomics features selected by the different selection methods.

From: Wavelet radiomics features from multiphase CT images for screening hepatocellular carcinoma: analysis and comparison

Classifiers

Feature selection models

No. selected/total features

\(\varvec{\text{F}_1}\) (95% CI)

AUC (95% CI)

LSR

Proposed logistic sparsity

29/2364 (1.23%)

0.89 (0.87–0.90)

0.96 (0.95–0.96)

Logistic ridge\(^{*,+}\)

796/2364 (33.67%)

0.84 (0.83–0.85)

0.92 (0.91–0.93)

Logistic elastic-net\(^{*,+}\)

290/2364 (12.27%)

0.81 (0.80–0.83)

0.90 (0.89–0.91)

LASSO20\(^{+}\)

30/2364 (1.27%)

0.89 (0.87–0.90)

0.95 (0.94–0.96)

PCA30\(^{*,+}\)

194/2364 (8.21%)

0.66 (0.64–0.68)

0.75 (0.73–0.77)

FVT & FCT20\(^{*,+}\)

887/2364 (37.52%)

0.73 (0.71–0.75)

0.82 (0.81–0.84)

Wrapper LR48\(^{*,+}\)

30/2364 (1.27%)

0.78 (0.76–0.79)

0.82 (0.81–0.84)

Wrapper RF60\(^{*,+}\)

30/2364 (1.27%)

0.82 (0.81–0.84)

0.88 (0.87–0.89)

MLP

Proposed logistic sparsity

29/2364 (1.23%)

0.87 ( 0.86–0.89)

0.94 ( 0.93–0.95)

Logistic ridge\(^{*,+}\)

796/2364 (33.67%)

0.82 ( 0.80–0.83)

0.90 ( 0.89–0.92)

Logistic elastic-net\(^{*,+}\)

290/2364 (12.27%)

0.82 ( 0.81–0.84)

0.91 ( 0.89–0.92)

LASSO20

30/2364 (1.27%)

0.87 ( 0.86–0.88)

0.94 ( 0.93–0.95)

PCA30\(^{*,+}\)

194/2364 (8.21%)

0.64 ( 0.62–0.66)

0.70 ( 0.68–0.72)

FVT & FCT20\(^{*,+}\)

887/2364 (37.52%)

0.77 ( 0.75–0.79)

0.84 ( 0.83–0.86)

Wrapper LR48\(^{*,+}\)

30/2364 (1.27%)

0.75 ( 0.74–0.77)

0.83 ( 0.82–0.85)

Wrapper RF60\(^{*,+}\)

30/2364 (1.27%)

0.82 ( 0.81–0.84)

0.89 ( 0.88–0.91)

SVM

Proposed logistic sparsity

29/2364 (1.23%)

0.89 ( 0.88–0.90)

0.96 ( 0.95–0.97)

Logistic ridge\(^{*,+}\)

796/2364 (33.67%)

0.83 ( 0.81–0.84)

0.91 ( 0.90–0.92)

Logistic elastic-net\(^{*,+}\)

290/2364 (12.27%)

0.77 ( 0.75–0.79)

0.86 ( 0.85–0.88)

LASSO20\(^{+}\)

30/2364 (1.27%)

0.88 ( 0.87–0.89)

0.95 ( 0.94–0.96)

PCA30\(^{*,+}\)

194/2364 (8.21%)

0.65 ( 0.63–0.67)

0.73 ( 0.71–0.75)

FVT & FCT20\(^{*,+}\)

887/2364 (37.52%)

0.71 ( 0.69–0.73)

0.79 ( 0.77–0.81)

Wrapper LR48\(^{*,+}\)

30/2364 (1.27%)

0.77 ( 0.76–0.79)

0.82 ( 0.80–0.83)

Wrapper RF60\(^{*,+}\)

30/2364 (1.27%)

0.82 ( 0.81–0.84)

0.88 ( 0.86–0.89)

  1. The methods marked with asterisk (\(*\)) and/or plus (+) symbols indicate statistical significance compared to the proposed LSR method using the same classifier at a confidence level of 95%, as determined by the t-test and/or DeLong’s test, respectively.
  2. The significant values, compared to the corresponding group in the first column, are in bolds.
Back to article page