ANNs performance during the best training epoch in large scale GA optimization.
Scheme of the best net architecture after >26 cycles of GA optimization.
Performance during training the epochs for the selected network. The 21th epoch resulted as the best one.
ROC curve after blind simulation with Db2 using large scale algorithm.
ROC curve after blind simulation with Db2 using SVM.
Radiation induced acute toxicity RTOG/EORTC classification criteria (Ref. 24).
List of the parameters used as ANN inputs.
Distribution of grade 0 and 1 in GI, GU and overall toxicity analysis (n = 321 patients).
Summary of candidate ANN architectures. [a,b,c] refers to ANNs with only one hidden layer. [d,e,f] refers to ANNs with two hidden layers.
Threshold based classification for the ROC curve of the ANN selected by large scale optimization.
Overall accuracy after blind simulation.
Article metrics loading...
Full text loading...