Table 6 13 features (*) are selected by both methods (RF and KO): gaitSpeed_Off, ABC, BMI, PIGD_score, cerebellum, X2.11, partII_sum, Attention, DGI, Tremor_score, FOG_Q, R_fusiform_gyrus, H_and_Y_OFF.

From: Model-based and Model-free Machine Learning Techniques for Diagnostic Prediction and Classification of Clinical Outcomes in Parkinson’s Disease

Random Forests

Knockoff

Features

Frequency

Features

Frequency

gaitSpeed_Off*

0.924

gender

0.917

ABC*

0.874

X2.11*

0.753

BMI*

0.824

ABC*

0.488

PIGD_score*

0.644

gaitSpeed_Off*

0.452

TUG_OFF

0.614

partII_sum*

0.425

cerebellum*

0.596

H_and_Y_OFF*

0.421

X2.11

0.568

cerebellum*

0.386

partII_sum*

0.522

PIGD_score*

0.359

brainstem

0.406

FOG_Q*

0.351

L_inferior_occipital_gyrus

0.402

X1.8

0.351

L_supramargiNAl_gyrus

0.402

BMI*

0.347

Attention*

0.392

X3.10gait_off

0.339

DGI*

0.378

DGI*

0.296

L_hippocampus

0.344

Attention*

0.296

L_fusiform_gyrus

0.342

R_fusiform_gyrus*

0.238

Tremor_score*

0.336

X2.13

0.226

FOG_Q*

0.328

X3.17d

0.211

R_fusiform_gyrus*

0.328

X4.3

0.187

R_parahippocampal_gyrus

0.318

Tremor_score*

0.176

H_and_Y_OFF*

0.308

X3.13

0.172

Back to article page