Clustering Methodology in Mixed Data Sets

Abstract

One of the most challenging tasks of data analysis is finding clusters in mixed data sets, as they have numerical and categorical variables, and lack a labeled variable to serve as a guide. These clusters could serve to summarize all the variables of a data set into one and be able to find information more easily than generating summarizations for each variable. In this research thesis, a methodology of clustering on mixed data sets is proposed, which yields better results than the methods applied in the state of the art.