Figure 4 | Scientific Reports

Figure 4

From: A priori prediction of breast cancer response to neoadjuvant chemotherapy using quantitative ultrasound, texture derivative and molecular subtype

Figure 4

Summary of classification performance with hold-out validation. Average tumour response classification results obtained from 10 hold-out data set using KNN and SVM-RBF algorithms. The top panel (a) shows data with 20% testing (80% training) and the bottom plot (b) with 10% testing (90% training) hold-out approaches. The classification performance with 90% training was approximately similar to those with leave-one-out approach with bias and variance errors of 21% and 3%, respectively.

Back to article page