CMStatistics 2018: Start Registration
View Submission - CMStatistics
Title: Multivariate dependence structures for ordinal data: A $\phi$-divergence based approach Authors:  Maria Kateri - RWTH Aachen University (Germany) [presenting]
Abstract: Dependence structures among ordinal variables will be studied in connection to $\phi$-divergence measures. Log-linear models for ordinal classification variables will be redefined through the Kullback-Leibler divergence and embedded in generalized families of models derived by replacing the Kullback-Leibler by the $\phi$-divergence. The scaling role of the $\phi$-divergence in constructing models for ordinal data and its effect on describing the underlying dependence structure will be discussed. The focus will be on high-dimensional contingency tables. Representative applications for members of the $\phi$-divergence based model families will be presented.