Skip to main content
European Radiology Experimental

Table 2 Performance of different feature combinations using the decision tree 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

{6}

73.2

69.1

0.652

{6, 1}

71.2

69.2

0.653

{6, 1, 2}

71.9

73.3

0.720

{6, 1, 2, 8}

70.6

75.0

0.744

  1. {1}: area difference with equivalent ellipse; {2}: orientation; {6}: area difference between the convex hull and tumour; {8}: entropy
Back to article page