Skip to main content
. 2020 Jan 17;10:607. doi: 10.1038/s41598-020-57549-y

Figure 2.

Figure 2

Recall vs. hourly false positive detections for generalization data from different geographic regions and dates. Results were averaged across 10 runs, each with random parameter initializations. The number of upcalls in the data on each of the six trial days is indicated by n.