Table 2 Features selected using (1) logistic regression (LR), (2) support vector machine (SVM), (3) random forest (RF), (4) elastic net (EN), and (5) relaxed linear separability method (RLS). For full feature names, see Table 1.

From: Balancing accuracy and cost in machine learning models for detecting medial vascular calcification in chronic kidney disease: a pilot study

Feature

LR

SVM

RF

EN

RLS

Cost (USD)

Sum

Age

X

X

X

X

X

0

5

Copeptin

X

X

X

X

X

58

5

Diabetes mellitus

X

X

 

X

X

2

4

Choline

 

X

X

X

X

9

4

Osteoprotegerin

 

X

X

X

X

5*

4

Sex, male

X

X

 

X

X

0**

4

BMI

 

X

X

X

 

0**

3

FBMI

X

X

  

X

0**

3

Sclerostin

 

X

X

X

 

5*

3

CTX

 

X

  

X

29

2

duMGP

 

X

 

X

 

5*

2

Homocysteine

 

X

 

X

 

23

2

IgMantiPC

 

X

 

X

 

7*

2

AGEAF

 

X

   

0**

1

Angiopoietin 2

 

X

   

5*

1

ApoB1

    

X

21

1

fT3

    

X

8

1

fPTG

    

X

5

1

GlaOC

    

X

23

1

GluOC

 

X

   

5*

1

hsCRP

 

X

   

7

1

IGF1

 

X

   

21

1

IgMantiMDA

 

X

   

23

1

LBMI

 

X

   

0**

1

PTX3

    

X

5*

1

TMAO

    

X

44

1

TroponinT

 

X

   

14

1

TSH

    

X

9

1

Uric Acid

    

X

9

1

Total number of features

5

21

6

11

16

  
  1. *Not measured in clinical practice. The cost is calculated based on the price of the kit per measurement.
  2. **Cost disregarded because of the relatively low machine expenses; the price per measurement is negligible when assuming testing of numerous individuals.
Back to article page