Modulo similarity in comparing histograms
Pasi Luukka, Mikael Collan · Advances in intelligent systems research/Advances in Intelligent Systems Research · 2015
Histograms are a tool for graphical representation of frequency data and thus helpful in creating a fast understanding of, e.g., contents of frequency data.Comparing histograms is topic of increasing importance due to an increase in the availability of data sets containing frequency information.Automatic data collection from "everywhere" has made collection of frequency data very common.As many different types of similarities exist, our focus is on Łukasiewicz logicbased similarity and we present two new measures, the "modulo similarity" measure and the "maximum pair assignment compatibility" measure.These measures do not use PDF conversion, or vector-based approaches, in the comparison of histograms, but concentrate on the data samples used to form histograms.We illustrate the usefulness of these measures with numerical examples.