Table 3 Variable selections model’s adjusted R2, AIC and BIC, NHANES 1999–2002 training set (n = 2965).

From: Development and validation of a prediction equation for body fat percentage from measured BMI: a supervised machine learning approach

Best 3 models

Selected variables

Adjusted R2

AIC

BIC

Model 1

BMI*Male, BMI*Age, BMI*Hispanic, BMI*Black, BMI* Low Education, BMI^2*Male, BMI^2*Age, BMI^2*Hispanic, BMI^2*Black, BMI^2*Low Education

0.8681

7100.0

7102.3

Model 2

BMI*Male, BMI*Age, BMI*Hispanic, BMI*Black, BMI^2*Male, BMI^2*age, BMI^2*Hispanic, BMI^2*Black, BMI^2*Low education

0.8680

7100.5

7102.7

Model 3

BMI*Male, BMI*Age, BMI*Hispanic, BMI*Black, BMI*Low Education, BMI^2*Male, BMI^2*Age, BMI^2*Hispanic, BMI^2*Black

0.8680

7101.5

7103.7

  1. The metrics for the top performing model is in Bold.
  2. Forced variables are BMI, BMI^2, gender, race, education, income, age.
  3. Variables available for selection: interaction terms between BMI and SES variables, and BMI squared and SES variables.
  4. Exit and entry levels: 0.2
  5. AIC: Akaike's Information Criteria, BIC: Bayesian Information Criteria.
Back to article page