Use machine learning to help identify possible sarcopenia cases in maintenance hemodialysis patients

BMC Nephrology

Table 2 Top 15 features ranked by the average ranking of feature importance and absolute feature weight of lasso regression

HGS Handgrip strength, FBG Fasting blood glucose, PTH Parathyroid hormone, pre-CRE Pre-dialysis creatinine, AG Anion gap, post-CRE Post-dialysis creatinine, AST Aspartate aminotransferase, LY% Lymphocyte percentage, TP Total protein, RDW-CV Red blood cell distribution width—coefficient of variation, CysC Cystatin C, s-Mg Serum magnesium, AWVA Arm without vascular access, WC Waist circumference, TBIL Total bilirubin, TG Triglyceride, AST/ALT Aspartate aminotransferase/alanine aminotransferase, SMI Skeletal muscle index, s-Fe Serum ferritin
^aImportance value calculated by RF
^bRanking of importance value calculated by RF
^cAbsolute feature weight of lasso regression
^dRanking of absolute feature weight of lasso regression
^eAverage ranking

ISSN: 1471-2369