Information theory for data processing

Applying information theory analysis for the solution of biomedical data processing problems. 

American Journal of Bioinformatics, 3 (1), 17-29, 2015

David Blokh and Ilia Stambler

Abstract

The use of information-theoretical methods can be highly valuable for the solution of biomedical data processing problems. Some of the problems that can be solved by those methods include: The assessment of the influence of diagnostic parameters, biomarkers and risk factors, on the emergence of disease; the discretization of diagnostic parameters; the analysis of a combined influence of a group of parameters; the partition of a group of diagnostic parameters according to the amount of diagnostic information contained in those parameters; the analysis of the parameters’ heterogeneity or variability and more. To illustrate the solution of those problems, we use a data base on diabetes patients. There are grounds to believe that an increasing application of information-theoretical methodologies in biomedical research will lead to significant practical dividends for diagnosis and therapy.

Full text:

http://thescipub.com/abstract/10.3844/ajbsp.2014.17.29

http://thescipub.com/PDF/ajbsp.2014.17.29.pdf