Performance comparison.

Performance values of multiple methods that were trained and tested on the same demographic model and selection coefficient. The performance values were found by computing the accuracy of the trained model on a balanced, held-out set. See Table C in S1 Text for additional models and summary statistics. These accuracies are also visualized in Figs D and E in S1 Text. Binomial-derived standard errors for these proportions are at most 0.5% for the single-population demographic model, and 1.1% for the three-population demographic model.