Skip to main content
. 2022 Aug 25;9(29):2202638. doi: 10.1002/advs.202202638

Table 1.

Performance metrics of the RF model for feature selection on the full‐feature dataset and after refining the dataset

Average F‐score Min (F1‐score) Max (F1‐score) StDev (F1‐score) Average OoB‐score
Full dataset Model a) 0.95
Printable 0.91 0.83 1.00 0.058
Refined dataset Model b) 0.96
Printable 0.94 0.91 1.00 0.039
Not‐Printable 0.99 0.98 1.00 0.007
a)

Data from 10 000 iterations on (n) repeated k‐fold cross‐validation (n = 20, k = 5) on full‐feature dataset.

b)

Data from (n) repeated k‐fold cross‐validation (n = 20, k = 5)