Table 3 Encoder network parameters (I: input channel size, O: output channel size, K: kernel size, S: stride size, P: padding size, R: (dropout) rate)
From: Predicting multiple sclerosis severity with multimodal deep neural networks
|  | Conv1d | Dropout | Conv1d | ReLU + Dropout | Conv1d | ReLU + Dropout | Pooling |
|---|---|---|---|---|---|---|---|
Channel 1 (Lab tests) | I: 1, O: 8, K: 7, S: 2 | R: 0.3 | I: 8, O: 8, K: 4, S: 2 | R: 0.3 | I: 8, O: 1, K: 3, S: 2 | R: 0.3 | Avg. |
Channel 2 (Vital Sign Observation) | I: 1, O: 8, K: 3, S: 2 | R: 0.3 | I: 8, O: 1, K: 2, S: 2 | R: 0.3 | Â | Â | Avg. |
Channel 3 (Medication) | I: 1, O: 8, K: 3, S: 2 | R: 0.3 | I: 8, O: 1, K: 2, S: 2 | R: 0.3 | Â | Â | Avg. |