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).
