Skip to main content

View full-text article in PMC

. 2017 Nov 22;12(11):e0188532. doi: 10.1371/journal.pone.0188532

Table 4. Feature importance for the top two performing algorithms for both triage and exacerbation models.

Triage		Exacerbation
GB	LR	GB	LR
Age	O₂Sat_(87, 89]	cough_3	cough_3
BMI	cough_3	sputum(col+vol)_0	shortbreath_3
Height	sputum(col+vol)_0	shortbreath_3	sputum(col+vol)_0
O₂Sat_(87, 89]	shortbreath_3	wheeze_3	wheeze_3
shortbreath_3	wheeze_3	sputum(col+vol)_1	sputum(col+vol)_1
Infection	Sleeplessness	Δfev1_(-15, -5]	Δfev_(0, 10]
Sleeplessness	Infection	Δfev1_(0, 10]	med_comp_3
cough_3	hr_(100, 110]	O₂Sat_(91, 93]	cough_1
wheeze_3	fev1_(0, 20]	sputum_vol_1	Δfev1_(-15, -5]
fev1_(0, 20]	O₂Sat_(85, 87]	fev1_(90, 100]	O₂Sat_(87, 89]
sputum(col+vol)_0	O₂Sat_(0, 85]	med_comp_3	Sleeplessness
hr_(100, 110]	Δfev1_(-100, -15]	Age	ΔO₂Sat_[-4,-2)_base < 93
O₂Sat_(85, 87]	hr_(110, 120]	shortbreath_1	O₂Sat_(85, 87]
heartrate_(110, 120]	sputum(col+vol)_1	Δfev_(-100, -15]	sputum_col_1
cough_1	hr>120	Sleeplessness	Δfev_(-5, 0]