Skip to main content

Table 2 Summary of balanced accuracies and weighted F1-scores of random forest models

From: Random forest algorithms to classify frailty and falling history in seniors using plantar pressure measurement insoles: a large-scale feasibility study

 

Frailty predictions

Fall predictions

Accuracy

F1-score

Accuracy

F1-score

Whole population

(N = 712)

0.75 ± 0.04

0.77 ± 0.03

0.57 ± 0.05

0.62 ± 0.04

Age-group:

≥65 years (N = 585)

0.76 ± 0.04

0.77 ± 0.04

0.55 ± 0.05

0.59 ± 0.04

65–69 years (N = 198)

0.68 ± 0.11

0.78 ± 0.05

0.60 ± 0.10

0.67 ± 0.06

70–74 years (N = 153)

0.68 ± 0.09

0.70 ± 0.07

0.49 ± 0.10

0.53 ± 0.08

≥ 75 years (N = 234)

0.71 ± 0.06

0.71 ± 0.06

0.54 ± 0.08

0.58 ± 0.07

Sex:

Women (≥ 65 years, N = 409)

0.72 ± 0.04

0.74 ± 0.04

0.56 ± 0.05

0.61 ± 0.04

Men (≥ 65 years, N = 176)

0.78 ± 0.07

0.79 ± 0.05

0.49 ± 0.09

0.56 ± 0.07

Frail state:

Frail (N = 203)

  

0.49 ± 0.08

0.53 ± 0.07

Non-frail (N = 509)

  

0.58 ± 0.06

0.64 ± 0.04

  1. Results are presented as mean ± standard deviation