Skip to main content
European Radiology Experimental

Table 6 Performance of different feature combinations using the support vector machine method

From: Performance of machine learning software to classify breast lesions using BI-RADS radiomic features on ultrasound images

Features

Sensitivity (%)

Specificity (%)

Area under the curve

{4}

64.3

80.5

0.746

{4, 2}

67.1

76.2

0.798

{4, 2, 10}

67.1

78.8

0.807

{4, 2, 10 ,8}

68.6

76.2

0.814

{4, 2, 10 ,8 ,1}

71.4

76.9

0.840

  1. {1}: difference area with equivalent ellipse; {2}: orientation; {4}: number of peaks on the distance vector; {8}: entropy; {10}: lesion size
Back to article page