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

Table 3

The pipeline of the model with the best performance.

Modeling steps	Method
Data normalization	Min-max
Dimension reduction	PCA
Feature selection	KW
Classifier	SVM

(PCA: principal component analysis; KW: Kruskal–Wallis test; SVM: support vector machine).