Table 1. Comparison of novelty fall detectors.

Algorithm	AUC mean(std)	SE	SP
kNN k = 1	0.9554 (0.0052)	0.907	0.905
kNN-sum k = 2	0.9548 (0.0052)	0.913	0.901
K-means + 1NN (K = 800)	0.9575 (0.0056)	0.929	0.890
One-Class SVM	0.9439 (0.0060)	0.881	0.890