Skip to main content
. 2024 Dec 23;11:e64578. doi: 10.2196/64578

Table 2. Model performances.

Model Splitting techniques
Group 5-fold cross-validation Leave-one-subject-out Chronological split Odd-even-split Random split
R² score MAE R² score MAE R² score MAE R² score MAE R² score MAE
Random chance (P value) –3.306
(N/A)a
0.920
(N/A)
–6.833
(N/A)
0.910
(N/A)
–2.364
(N/A)
0.941
(N/A)
–2.115
(N/A)
0.920(N/A) –2.205
(N/A)
0.890(N/A)
Dummy regression (median) –0.289
(.79)
0.499
(.41)
–3.624
(.92)
0.557
(.72)
–0.107
(.99)
0.491
(.99)
–0.001
(.35)
0.482
(.48)
–0.007
(.72)
0.488
(.84)
Random forest regression (P value) –0.102

(.09)
0.455
(.04)
–4.392
(.99)
0.540
(.29)
–0.213
(.65)
0.519
(.81)
0.336
(<.001)
0.381
(<.001)
0.305
(<.001)
0.396
(<.001)
Linear regression (P value) –25.508
(.67)
0.588
(.50)
–37.258
(.71)
0.602
(.31)
–0.364
(.15)
0.534
(.18)
–0.179
(<.001)
0.445
(<.001)
–0.558
(.06)
0.459
(<.001)
Support vector regression (P value) –0.136
(.008)
0.468
(.004)
–4.006
(.88)
0.570
(.89)
–0.106
(.59)
0.439
(.87)
0.313
(<.001)
0.388
(<.001)
0.293
(<.001)
0.401
(<.001)
XGBoost regression (P value) –0.093
(.07)
0.455
(.03)
–3.568
(.41)
0.550
(.03)
0.084
(.98)
0.442
(.99)
0.339
(<.001)
0.380
(<.001)
0.289
(<.001)
0.399
(<.001)
a

N/A: not available.