The comparison of multi-dimensional data simalarity measurements on the recongnition of human activities

Abstract

Recently a tendency dominates that people acquire various technological equipment to observe their physical activities. In order to get information from the data collected by this equipment, it is necessary to know how to compare or classify it. The similarity measurements are used for the comparison of objects, however, different measurements evaluate different attributes of objects and it is often hard to determine which of them is the most accurate. In this article, the influence of similarity measure analysis is introduced, in order to find out how the similarity measurements impact the time series segment classification. The object of the research is which similarity measurements allow to determine the type of physiological activity the most accurately, when evaluating time series that define human physiological attributes and the attributes of their position in a space.