Table 1 Performance of the eight ML model for predicting sarcopenia on the test set

Full model-2-year

 LR

0.117

0.792 (0.739–0.837)

0.860

0.865

0.104

 DT

0.154

0.746 (0.683–0.787)

0.841

0.845

0.109

 SVM

0.141

0.734 (0.672–0.795)

0.864

0.864

0.109

 RF

0.123

0.790 (0.745–0.829)

0.853

0.863

0.105

 AdaBoost

0.402

0.784 (0.730–0.827)

0.841

0.839

0.176

 XGBoost

0.114

0.804 (0.750–0.853)

0.872

0.875

0.100

 LightGBM

0.063

0.778 (0.732–0.822)

0.861

0.865

0.106

 ANN

0.111

0.779 (0.730–0.817)

0.863

0.869

0.105

Full model-4-year

 LR

0.616

0.779 (0.736–0.820)

0.573

0.981

0.281

 DT

0.222

0.750 (0.699–0.802)

0.731

0.919

0.189

 SVM

0.366

0.786 (0.741–0.829)

0.683

0.953

0.121

 RF

0.498

0.782 (0.701–0.832)

0.676

0.958

0.188

 AdaBoost

0.500

0.749 (0.706–0.814)

0.789

0.915

0.239

 XGBoost

0.072

0.793 (0.759–0.837)

0.807

0.903

0.129

 LightGBM

0.010

0.795 (0.758–0.835)

0.845

0.882

0.121

 ANN

0.009

0.730 (0.665–0.778)

0.777

0.885

0.174

Modifiable model-2-year

 LR

0.111

0.783 (0.736–0.815)

0.865

0.869

0.104

 DT

0.224

0.756 (0.715–0.799)

0.854

0.862

0.112

 SVM

0.366

0.786 (0.741–0.829)

0.683

0.953

0.184

 RF

0.129

0.786 (0.751–0.833)

0.833

0.813

0.105

 AdaBoost

0.441

0.786 (0.750–0.832)

0.866

0.863

0.202

 XGBoost

0.069

0.795 (0.757–0.827)

0.859

0.864

0.106

 LightGBM

0.066

0.766 (0.723–0.813)

0.860

0.865

0.109

 ANN

0.089

0.787 (0.754–0.840)

0.863

0.861

0.104

Modifiable model-4-year

 LR

0.427

0.759 (0.707–0.804)

0.718

0.932

0.187

 DT

0.143

0.638 (0.560–0.702)

0.725

0.896

0.227

 SVM

0.123

0.685 (0.633–0.734)

0.844

0.862

0.119

 RF

0.237

0.768 (0.713–0.818)

0.802

0.888

0.131

 AdaBoost

0.499

0.728 (0.661–0.776)

0.768

0.897

0.291

 XGBoost

0.232

0.759 (0.711–0.808)

0.756

0.911

0.155

 LightGBM

0.074

0.769 (0.715–0.812)

0.834

0.877

0.130

 ANN

0.026

0.706 (0.642–0.765)

0.742

0.892

0.207