Fig. 2
From: A k-mer grammar analysis to uncover maize regulatory architecture
Comparison between models of the precision-recall curves. Comparison of models performance under balanced (a - b) and unbalanced holdout sets (c - d). For each model (k=8), the precision recall (PR) curve for all the regulatory datasets are shown, and the corresponding curves for classification of the same holdout set with a collection of PWMs (dotted lines). The PR curve shows the trade-off between precision and recall for different decision threshold. A high area under the curve represents both high recall (low false negative rate) and high precision (low false positive rate)