Abstract
The subject of features normalization plays an important central role in datarepresentation, characterization, visualization, analysis, comparison,classification, and modeling, as it can substantially influence and beinfluenced by all of these activities and respective aspects. The selection ofan appropriate normalization method needs to take into account the type andcharacteristics of the involved features, the methods to be used subsequentlyfor the just mentioned data processing, as well as the specific questions beingconsidered. After briefly considering how normalization constitutes one of themany interrelated parts typically involved in data analysis and modeling, thepresent work addressed the important issue of feature normalization from theperspective of uniform and proportional (right skewed) features and comparisonoperations. More general right skewed features are also considered in anapproximated manner. Several concepts, properties, and results are describedand discussed, including the description of a duality relationship betweenuniform and proportional feature spaces and respective comparisons, specifyingconditions for consistency between comparisons in each of the two domains. Twonormalization possibilities based on non-centralized dispersion of features arealso presented, and also described is a modified version of the Jaccardsimilarity index which incorporates intrinsically normalization. Preliminaryexperiments are presented in order to illustrate the developed concepts andmethods.