Fig. 2
From: Toward a hemorrhagic trauma severity score: fusing five physiological biomarkers
Evaluation of the mean test and mean train accuracy versus the number of training samples for the training size varied from 30-80 in steps of 5. Accuracy vs. training dataset size for a SVM-L, b EBDT, c ANN:BR, and d Evaluation of the influence of the size of the training set, expressed as a  % of available data, on the performance of the ANN:BR as expressed in the Mean-Square-Error for maximum epochs of 100. For unsorted data, trained ANNs were tested with a sliding window of 25 validation data set. For sorted data, trained ANNs were tested with a sliding window of 20 validation data set