This is a university project that allowed us to use some nonparametric statistics techniques on a real problem and develop a web application achieving our goal.
Our project purpose is to create a web page which evaluates the fairness of health insurance premium for US citizens, based on physical characteristics and lifestyle habits. Firstly, we analyze our dataset variables, investigating which ones are the most significative for our aim using nonparametric tests, such as ANOVA and permutation tests. Secondly, we build the best fitting model with Generalized Additive Models (GAMs), conducting a stepwise procedure, and we make premium predictions. Lastly, we implement the web application with the R package “Shiny” to assess prices for our web audience.
This project is entirely developed in R language.