Skip to main content
. Author manuscript; available in PMC: 2019 Dec 1.
Published in final edited form as: Smart Health (Amst). 2018 Jul 7;9-10:287–296. doi: 10.1016/j.smhl.2018.07.002

Table 2.

Performance comparison between the proposed algorithm and the vanilla AdaBoost

AdaBoost with Reject Option Vanilla AdaBoost
131 weak learners 218 weak learners 131 weak learners 218 weak learners
Number of Samples 2318
Valid / Invalid Alarms 549 / 1769
Sensitivity 100.00% 99.27% 95.26% 93.44%
Specificity 9.38% 23.12% 38.78% 43.81%
Silenced Alarms 166 413 712 813
Total False Negative 0 4 26 36