Statistical Measures for Contrast Patterns
James E. Bailey · 2012
Department of Computing and Information Systems, The University of Melbourne 2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 2.1.1 Terminology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14 2.2 Measures for Assessing Quality of Discrete Contrast Patterns . . 15 2.3 Measures for Assessing Quality of Continuous Valued Contrast Patterns . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18 2.4 Feature Construction and Selection: PCA and Discriminative Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19 2.5 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20 An important task when working with contrast patterns is the assessment of their quality or discriminative ability. In this chapter, we review a range of measures that may be used to assess the discriminative ability of contrast patterns. Some of these measures have their origins in association rules, others in statistics, and others in subgroup discovery. Our presentation is not exhaustive, since dozens of measures exist. Instead we present a selection that covers a number of the main types.