Figure 3

(a) Flow diagram illustrating the procedure to split the dataset and evaluate the model performance to classify ovarian cancer as responders or non-responders to carboplatin based chemotherapy. The tissue microarray (TMA) dataset contains cores representative of ovarian cancer surgical specimens from patients undergoing cytoreductive surgery followed by neoadjuvant chemotherapy. Data was divided into a training set (nā=ā40) used to select and train the model and a left out validation set (nā=ā10) to test the accuracy of the final model to predict chemo-sensitivity. (b) Accuracy of the model generated with each classifier.