Developing a system that can automatically detect health changes using transfer times of older adults

Table 9 Comparison of the results on transitions to gait models with varying parameters

Changing parameters	Scenario^a	FPR^b	DR^c	ARL^d
μ and σ	V _SU	0.20±0.07	100 %	7.70±3.34
	V _US	0.17±0.07	100 %	10.95±3.59
	V _SUS	0.20±0.08	100 %	10.40±4.50
	V _USU	0.20±0.08	100 %	9.05±3.33
μ only	V _{S T2}	0.20±0.09	100 %	7.85±2.94
	V _{U T1}	0.16±0.06	100 %	9.55±3.78
	V _{S T2S}	0.23±0.04	100 %	8.05±3.41
	V _{U T1U}	0.21±0.07	97.50 %	9.97±3.59
σ only	V _{S T1}	0.13±0.08	50 %	14.50±6.96
	V _{U T2}	0.09±0.07	25 %	7.20±9.47
	V _{S T1S}	0.15±0.06	35 %	11.57±9.07
	V _{U T2U}	0.11±0.09	37.50 %	12.87±8.25

This change in gait model parameters is defined as either a change in both μ, σ or either μ or σ (all transitions have a length of four weeks)
Notes
^aSimulation scenarios as described in Table 4
When two trends are present in the data
an average DR and ARL is calculated
^bAverage number of False Alerts per week
^cDetection Rate
^dAverage Run Length

ISSN: 1471-2288