Title :
Entropy Inequalities for the generalized Gaussian
Author :
Kitsos, Christos P. ; Toulias, Thomas L.
Author_Institution :
Dept. of Math., Technol. Educ. Inst. of Athens, Athens, Greece
Abstract :
The target of this paper is to discuss the existent Poincaré and Logarithm Sobolev Inequalities (PI and LSI resp.) for the Gaussian (normal) distribution which is essential in theoretical Statistics and plays an important role in Information Theory and Statistics. The adopted Mathematical backround is usually simplified in practical applications. The entropy, energy and variance are related through some order due to PI and LSI. The extended multivariate normal, being a generalized Gaussian, also obeys to LSI.
Keywords :
Gaussian processes; Poincare mapping; Poincaré existent; entropy inequalities; generalized Gaussian; logarithm Sobolev inequalities; mathematical backround; theoretical statistics; Channel capacity; Covariance matrix; Entropy; Gaussian distribution; Large scale integration; Space technology; Entropy power; Information measures; Logarithmic Sobolev Inequalities; Poincaré Inequalities;
Conference_Titel :
Information Technology Interfaces (ITI), 2010 32nd International Conference on
Conference_Location :
Cavtat/Dubrovnik
Print_ISBN :
978-1-4244-5732-8
