Fig. 4

Receiver operating characteristic (ROC) curves of classifiers for distinguishing pCR and non-pCR patients. a The classifier based on global chromatin (5 Mb windows) in Random Forest. b The classifier based on sub-compartments in Random Forest. c The classifier based on promoter profiles in Random Forest. d The classifier based on global chromatin (5 Mb windows) in Logistic Regression. e The classifier based on sub-compartments in Logistic Regression. f The classifier based on promoter profiles in Logistic Regression. g The classifier based on global chromatin (5 Mb windows) in Support Vector Machines. h The classifier based on sub-compartments in Support Vector Machines. i The classifier based on promoter profiles in Support Vector Machines. pCR, pathological complete response; non-pCR, non-pathological complete response; RF: Random Forest; LR: Logistic Regression; SVM, Support Vector Machines