$ ⌘K
// constructs / forecasting_accuracy_auc

Forecasting Accuracy (AUC)

id: forecasting_accuracy_auc

Area under the ROC curve for coup prediction models. Most published models achieve AUC of 0.75-0.85 but with high false positive rates given the low base rate. Ward et al. (2010) argue that in-sample statistical significance is insufficient and out-of-sample predictive accuracy should be the standard for evaluating coup models.

// usage

Used in 1 pax

direction not specified 0 findings