Table 3 Classification performance

# of Features

40,112

290

27

24

10

4

Random forest classifier (T0 dataset, leave-one-out cross-validation)

LOOCV accuracy

0.559 ± 0.497

0.618 ± 0.486

0.853 ± 0.354

0.912 ± 0.284

0.912 ± 0.284

0.794 ± 0.404

LOOCV precision

0.559 ± 0.497

0.618 ± 0.486

0.647 ± 0.478

0.647 ± 0.478

0.647 ± 0.478

0.618 ± 0.486

LOOCV recall

0.559 ± 0.497

0.618 ± 0.486

0.647 ± 0.478

0.647 ± 0.478

0.647 ± 0.478

0.618 ± 0.486

LOOCV F1 score

0.559 ± 0.497

0.618 ± 0.486

0.647 ± 0.478

0.647 ± 0.478

0.647 ± 0.478

0.618 ± 0.486

Random forest classifier (T0 training, self-test)

Training accuracy

1

1

1

1

1

1

Training precision

1

1

1

1

1

1

Training recall

1

1

1

1

1

1

Training F1 score

1

1

1

1

1

1

Random forest classifier (T0 training, T2 testing)

Test accuracy

0.647

0.647

0.706

0.706

0.735

0.765

Test precision

0.647

0.647

0.688

0.700

0.710

0.769

Test recall

1

1

1

0.955

1

0.909

Test F1 score

0.786

0.786

0.815

0.808

0.83

0.833