Prediction
Overview
This section covers the modules used for prediction in aedesproject-uif. These modules handle the generation of forecasts and risk scores based on the trained models. The model pickles of the final models from the ml.risk_model_dev module should be present in the model folder.
generate_forecasts
Description
This module generates forecasts by applying trained models on new data.
Main Function
Function Name
forecast_data(iso_country_code, adm, target, n_ar, freq)
Description
Main function used for generating forecasts.
Parameters
iso_country_code: ISO code of a country.adm: Subregion for which to generate forecasts.target: Target outcome to be forecasted.n_ar: Number of AR to be included in the initialization of the forecaster. This should be equal to the number used in model training.freq: Frequency of the dates in the training data. Refer to this link for the valid values.
Usage
from aedesproject_uif.predict.generate_forecasts import forecast_data
forecast_data("PHL", "Zamboanga del Norte", "cases", 12, "W")
generate_risk_scores
Description
This module generates hazard, vulnerability, lack of coping capacity, and consolidated risk scores in the aedesproject-uif.
Main Function
Function Name
risk_score_data(ISO_COUNTRY_CODE)
Description
Main function used for generating the individual and consolidated risk scores.
Parameters
ISO_COUNTRY_CODE: ISO code of a country.
Usage
from aedesproject_uif.predict.generate_risk_scores import risk_score_data
risk_score_data('PHL')