Skip to main content

Table 3 Permutation importance of the 20 most important features

From: Interpretable prediction of 3-year all-cause mortality in patients with chronic heart failure based on machine learning

Weight

Feature

0.1160 ± 0.0183

Number_Of_Hospitalizations

0.0340 ± 0.0239

Age

0.0298 ± 0.0282

Glomerular_Filtration_Rate

0.0234 ± 0.0159

BNP

0.0191 ± 0.0159

NYHA_Cardiac_Function_Classification

0.0128 ± 0.0144

Lymphocyte_Absolute_Value

0.0106 ± 0.0095

Serum_Albumin

0.0096 ± 0.0264

Hemoglobin

0.0096 ± 0.0124

Total_Cholesterol

0.0085 ± 0.0052

Pulmonary_Artery_Systolic_Pressure

0.0085 ± 0.0128

Low_Density_Lipoprotein_Cholesterol

0.0064 ± 0.0156

Absolute_Neutrophil_Value

0.0064 ± 0.0124

Serum_Sodium

0.0064 ± 0.0124

Serum_Uric_Acid

0.0053 ± 0.0117

Eosinophil_Absolute_Value

0.0053 ± 0.0223

Body_Mass_Index

0.0053 ± 0.0067

Left_Ventricular_Ejection_Fraction

0.0043 ± 0.0141

Albumin/Globulin

0.0043 ± 0.0080

Triglyceride

0.0032 ± 0.0185

Heart_Rate