Table 11 2D River problem: training, validation and testing dataset
From: Deep convolutional architectures for extrapolative forecasts in time-dependent flow problems
Dataset | Samples | Input | Output |
|---|---|---|---|
Training | 1 | \([z^{1},..., z^{n_t-1}, z^{n_t}]\) | \(z^{n_t+1}\) |
2 | \([z^{2},..., z^{n_t}, z^{n_t+1}]\) | \(z^{n_t+2}\) | |
... | ... | ... | |
50 | \([z^{50},..., z^{n_t+48}, z^{n_t+49}]\) | \(z^{n_t+50}\) (training end) | |
Validation | 51 | \([z^{51},..., z^{n_t+49}, z^{n_t+50}]\) | \(z^{n_t+51}\) |
... | ... | ... | |
60 | \([z^{60},..., z^{n_t+58}, z^{n_t+59}]\) | \(z^{n_t+60}\) | |
Testing | 1 | \([z^{1},..., z^{n_t-1}, z^{n_t}]\) | \([z^{n_t+1},..., z^{99}, z^{100}]\) |