246 – Important Issues in Clinical Trials, Meta-Analysis, and Diagnostic Medicine
Core/Non-Core Configuration Within Big Data Analytics
Turkan K. Gardenier
Pragmatica Corp.
John S. Gardenier
National Center for Health Statistics (Retired)
The authors expand on a Trinary or Triomial-based approach presented in two preceding annual Joint Statistical Meetings, tailoring the procedure to data streaming into Big Data. Core is defined as the central (0) region, with two adjoining Non-core (+) and (-) regions. Analytic issues related to streaming data, as contrasted with data analyzed as a whole,are summarized. Characteristics of Trinary (+/0/-) three-category approach are presented and contrasted with two-category or Binomial formulations. Two applications are included: (a) air quality monitoring for Nitrogen Dioxide spanning 20 years (1991 - 2010) in two cities in the US, and (b) lung cancer mortality data as examples of a vast array of significant issues in environment, genomics, economics, energy, public healthy and education which can benefit from this exploratory approach.