Skip to main content
. 2010 Jun 1;26(12):i168–i174. doi: 10.1093/bioinformatics/btq189

Fig. 4.

Fig. 4.

ROC curves for identifying periodic signals from four datasets of (A) 10 000 stationary periodic signals and 10 000 white noise signals, (B) 10 000 non-stationary periodic signals and 10 000 white noise signals, (C) 10 000 stationary periodic signals and 10 000 AR(1)-based random signals, and (D) 10 000 non-stationary periodic signals and 10 000 AR(1)-based random signals. Greater area under the ROC curve indicates better performance of the algorithm. ARSER gave the fewest false positives and false negatives compared with COSOPT and Fisher's G-test in all cases.