Figure 5From: Forecasting the locational dynamics of transnational terrorism: a network analytic approachOne-year-ahead forecast AUC for various models. To compute the forecast AUC, we use the parameters estimated by modeling Nt−1 based on N t 0 − 1 , t − 2 to forecast Nt based on N t 0 , t − 1 . The “Best” model is the one that has had the highest average predictive log score up to t. “Memory 1” only has a memory effect from the previous year and an edges term. The “Memory 5” model only has a memory effect from the previous 5 years and an edges effect.Back to article page