Identification of the main cluster in the scatter diagram of electoral data

Abstract

I describe and formalize a method for reconstructing falsified election results based on the analysis of a scatter diagram of polling stations in the space of electoral characteristics. I compare the results of applying this method with the reconstruction based on the analysis of the distribution of voters by electoral characteristics. This allows not only to identify and eliminate specific situations where one of the methods is ineffective, but also to obtain independent confirmation of the assumptions underlying each of them.