Sentence examples for false alarm range from inspiring English sources
All multiple comparisons were significant after adjustment for repeated testing for the false alarm range 0 0.005 (all p ≤ 0.005).
Comparison of the weighted AUC analyses for the seven cusum algorithms found no significant difference for the 0 0.005 false alarm range (Friedman χ2 = 10.7, p = 0.10).
However, no multiple comparisons were significant after adjustment for repeated testing for the false alarm ranges 0 1 (all 0.17 ≤ p ≤ 0.85), and 0 0.01 (C1 all p ≤ 0.02, C2 all p ≤ 0.08, C3 all p ≤ 1.00).
Comparison of the AUC analyses for the seven cusum algorithms found a significant difference for the 0 1, 0 0.01, and the 0 0.005 false alarm ranges (Friedman χ2 = 70.3, p < 0.0001, Friedman χ2 = 70.6, p < 0.0001 and Friedman χ2 = 61.9, p < 0.0001 respectively, Table 3).
With the exception of the comparison of EARS C3 and the 7-day negative binomial cusum (p = 0.01 and p = 0.04 respectively), all multiple comparisons were significant after adjustment for repeated testing for the false alarm ranges 0 1 and 0 0.01 (all p ≤ 0.002).
Detection threshold was empirically calculated by considering the normalized histogram of the test statistic under H 0 and corresponds to probability of false alarm of range 0 to 1.
For train/dev data, it is clear that the Fusion (DNN) + Posteriorgram system performs similar to the Fusion (DNN) system for most of the range, except when false alarm rate is low, where the Fusion (DNN) + Posteriorgram system performs better, and that the DNN-based fused systems outperform the SGMM-based systems counterpart in general.
Data for three participants with excessive false alarm rates (ranging 0.30 0.67) and for one participant with excessive miss rates (as high as 0.79 in the right-single target condition) were excluded.
Because the maximum false alarm rate is limited by the lineup size, we actually estimated the pAUC for the showup and six-person lineup ROCs (those compared visually by Lampinen) using a false alarm rate range of 0 to either 0.10 (only the highest-confidence responses) or 0 0.16 (essentially the full ROC for the six-person lineup).
constant false alarm rate.
