Mendes, Jorge MoraisMattos, Acácio Luis Ramalho2022-01-042022-01-042021-12-13http://hdl.handle.net/10362/130221Dissertation presented as the partial requirement for obtaining a Master's degree in Statistics and Information Management, specialization in Information Analysis and ManagementUsing nominal variables with many levels as predictors in statistical modelling can be a difficult task. Ways to group the levels of such variables optimally becomes of special interest. In this work we propose a procedure and a tool implemented in R language to deal with this type of situation, in order to create an optimally-recategorized variable to modelling or simply for descriptive purposes. An example with a real data in Insurance environment will be conducted to illustrate the performance, as well the advantages and disadvantages.engOptimal Level collapsingLevel groupingNominal variablesCategorical variablesRelevelingolcmaster thesis202839311