In this study we develop machine learning (ML) models to predict houshold income and body mass index (BMI) based on demographic data such as participant age, gender, ethnicity, education level using the NHANES public dataset.
The demographic data is in DEMO_I.xlsx and description of the demographic data can be found in DEMO_I.pdf The body measurements data is in BMX_I.xlsx and description of the different variables in the data can be found in BMX_I.pdf
The following steps are involved in this study:
- Preliminary exploration of data using univariate and bivariate analysis techniques
- Feature selection and feature transformation
- Model development and validation
- Prediction on test set to evaluate performance of model
The ML models investigated include:
- Random Forest
- Gradient Boosting
- SVM
- K-Nearest Neighbor
- Logistic Regression
- Naive Bayes
- MLP Regressor
- Linear Regression, Ridge Regression, Lasso Regression