Table 6 The optimal hyperparameters and their configurations for different models.
From: Exogenous variable driven deep learning models for improved price forecasting of TOP crops in India
Models | Hyperparameters | Values |
|---|---|---|
ANN | No. of hidden layers | 1 |
No. of neurons | 21 | |
Activation function | ReLU | |
SVR | Kernel | RBF |
c | 0.2 | |
\(\varepsilon\) | 0.01 | |
RFR | No. of trees | 500 |
Maximum features | 10 | |
Minimum No. of samples to split | 2 | |
XGBoost | No. of trees | 800 |
Maximum depth of tree | 5 | |
Boosting type | Tree boosting | |
NBEATSX | Fully connected layers | 4 |
Lookback | 7 | |
Horizon | 1 | |
Stacks | 30 | |
Neurons per layer | 512 | |
Epochs | 500 | |
Loss function | MAE | |
Optimizer | Adam | |
TransformerX | No. of layers | 4 |
Embedding dimension | 128 | |
No. of heads | 8 | |
No. of neurons | 512 | |
Dropout rate | 0.1 |
