Fig. 3From: Optimal length of R–R interval segment window for Lorenz plot detection of paroxysmal atrial fibrillation by machine learningRelationships between classification performance metrics and LP segment window length in the test data from the ALLSTAR database. The LP images were annotated with the strict criteria. Panels a and b show sensitivity, specificity, positive and negative predictive values (PPV and NPV), accuracy, and area under the curve (AUC). Panels c and d show positive and negative likely hood ratios, respectively. The x-axis in b–d is the common log values of segment window length in beat. Dotted lines represent regression lines with parabolic functionsBack to article page