OVERVIEW OF THE PREDICTION MODEL
The fatalities model
Model version: 001 (002 coming soon)
PREDICTED TYPE(S) OF VIOLENCE
Dichotomous outcome: Predicted probability of at least 25 battle-related deaths (BRDs) per country-month and at least 1 BRD per PRIO-GRID-month.
Read the technical report presenting the first version of the fatalities model (fatalities001)
ABOUT THE INPUT DATA
The features informing the VIEWS models
Key feature sets
Political institutions, democracy
Climate & societal vulnerability
Natural and social geography
Food security and access to basic needs
GENERATING THE FORECASTS
Model training procedures
Sub-models: combinations of feature sets and algorithms
Ensemble models: groups of sub-models using “the wisdom of the crowd”
- The country-level ensemble model combines the predictions from each of the sub-models using a genetic algorithm that assigns different weights to the contribution from each model in order to maximise predictive performance.
- The sub-national ensemble model, in turn, uses a simple unweighted average of the sub-model results.