Is it possible to get deepnet to output the prediction probabilities for each sample in a test dataset for a given model's parameters? I am doing some ROC analysis and it would be great if I could get the probability for each class for each sample. Then, I could just vary the p-value threshold to generate my ROC curves. Does anyone know of a way to do this with deepnet?
Is it possible to get deepnet to output the prediction probabilities for each sample in a test dataset for a given model's parameters? I am doing some ROC analysis and it would be great if I could get the probability for each class for each sample. Then, I could just vary the p-value threshold to generate my ROC curves. Does anyone know of a way to do this with deepnet?