Skip to main content

Changing the Forecasting Model

Author: Cansu Moran

Follow these steps to update the forecasting model in the project:

1. Locate the Forecasting Module​

  • Identify the main file handling the current forecasting model, forecast_model.py under forecast directory.

2. Add the New Forecasting Model​

  • Install any dependencies required for the new model (e.g., via pip): 
    pip install <new-model-library>
  • Create a new file for the model or modify the existing one: 
    /forecast/<new_model_name>.py
  • Implement the new model logic. Ensure it adheres to the expected input and output format used in the project.

3. Update the Model Selection Logic​

  • Update updatePredictions() function in index.py under forecast directory so that the data is passed to your selected model. 
    fcs_pred, rcsi_pred = new_forecast_model.predict(data)

Key Points:​

  • fcsGraph contains data points for Food Consumption Score (FCS), including upper and lower bounds.
  • rcsiGraph contains data points for Reduced Coping Strategies Index (rCSI), also with upper and lower bounds.

Your new model's function should be able to handle this data format.

4: Ensure Proper Output Format​

The function should return predictions as fcs_pred and rcsi_pred. The values are retrieved from the prediction results in the following structure:

{
    'fcs': round(fcs_pred[i][0]),
    'fcs_high': round(fcs_pred[i][1]),
    'fcs_low': round(fcs_pred[i][2]),
    'rcsi': round(rcsi_pred[i][0]),
    'rcsi_high': round(rcsi_pred[i][1]),
    'rcsi_low': round(rcsi_pred[i][2])
}