Model Details The model is configured with the following architecture parameters:
Input Dim: Number of input features Hidden Dim: 50 (Number of hidden units in each LSTM layer) Output Dim: 1 (The predicted severity score) Forecast Horizon: Number of future steps the model predicts (e.g., 1 step ahead) Layers: 2 LSTM layers Dropout Rate: 0.3 (For regularization) Example Data The model expects input data in the following format:
Feature Description Total Population Total population of the affected area Landmass Affected Area affected by the crisis (in square km) People Exposed Number of people exposed to the crisis (Other features) Additional relevant crisis features License This model is available under the MIT License (or your preferred license).
- Downloads last month
- 4