Skip to main content

Table 4 Feature importances obtained by the Boruta algorithm in arbitrary units

From: Machine learning-based prediction of fainting during blood donations using donor properties and weather data as features

 

Relative feature importance

Systolic blood pressure

41 (38–44)

Diastolic blood pressure

38 (30–45)

Ambient temperature

36 (34–39)

Relative humidity

33 (31–36)

Dew point

33 (30–35)

Atmospheric pressure

33 (31–35)

Percentage sunshine

28 (25–30)

Peek wind speed data

27 (25–30)

Peek wind direction data

25 (25–29)

Gender

23 (20–25)

Weight

23 (20–25)

BMI

23 (19–26)

Height

21 (20–22)

Wind direction data

12 (10–15)

Body temperature

10 (10–11)

  1. Mean and 95% interval are given