New weighed similarity indexes for market segmentation using categorical variables

In this paper we introduce new similarity indexes forcategorical data with nominal scale. In contrast to traditionallyused similarity measures, they also consider the frequency of themodalities of each attribute in the sample. This feature is usefulwhen dealing with rare categories, since it makes sense todifferently evaluate the pairwise presence of a rare category fromthe pairwise presence of a widespread one. We also propose aspecific weighted index for dependent categorical variables. Thesuitability of the proposed measures from a marketing researchperspective is shown using two real world data sets.