Radiomics Analysis Based on Automatic Image Segmentation of DCE-MRI for Predicting Triple-Negative and Nontriple-Negative Breast Cancer

Table 4

The selected features for the model according to validation performance.

Features	Coefficient in model
PCA_feature_1	0.932
PCA_feature_2	2.886
PCA_feature_4	-1.020
PCA_feature_7	0.329
PCA_feature_11	0.597
PCA_feature_15	1.014
PCA_feature_23	1.449
PCA_feature_24	0.980
PCA_feature_34	-1.338
PCA_feature_37	1.830
PCA_feature_39	-1.238
PCA_feature_41	1.412
PCA_feature_44	1.174
PCA_feature_46	0.932
PCA_feature_52	-0.897

(PCA: principal component analysis.).