Skip to main content
. 2018 Jul 27;15(8):1596. doi: 10.3390/ijerph15081596

Figure 6.

Figure 6

Ten models with the lowest RMSEs for each infectious disease. The numbers in parentheses indicate each model’s optimizer, activation, and number of epochs, respectively. (optimizer) 1: Adadelta, 2: Adagrad, 3: Adam, 4: Adamax, 5: Nadam, 6: RMSProp, and 7: SGD, (activation function) 1: ELU, 2: ReLU, 3: SELU, and 4: SoftPlus, (number of epochs) 1: 400, 2: 600, 3: 800, and 4: 1000. (a) The top-10 deep learning models and traditional analysis models with the lowest RMSEs in chickenpox, (b) The top-10 deep learning models and traditional analysis models with the lowest RMSEs in scarlet fever, (c) The top-10 deep learning models and traditional analysis models with the lowest RMSEs in malaria.