142nd APHA Annual Meeting and Exposition

Annual Meeting Recordings are now available for purchase

308199
A Study of U.S. Obesity Rate and Related Risk Factors Based on Support Vector Machine and Partial Least Squares Regression Models

142nd APHA Annual Meeting and Exposition (November 15 - November 19, 2014): http://www.apha.org/events-and-meetings/annual
Tuesday, November 18, 2014

Chau-Kuang Chen, EDD , School of Graduate Studies, Meharry Medical College, Nashville, TN
Background: Data from the CDC has shown that the dramatic rise in obesity began in the 1990s when there was less than 10 to 14 percent.  A decade later, there was a recorded 10 percent increase in the prevalence of obesity range.  The objective of this study was to identify important risk factors contributing to obesity rate within all 50 states based on the relative importance and variable importance in the projection (VIP).  

Method: The 2000-2011 U.S. obesity data for all 50 states and related risk factors were extracted from various sources including: United States Department of Labor, Centers for Disease Control and Prevention, and United States Census Bureau. The risk factors included per capita income, unemployment rate, poverty rate, alcohol consumption, and physical exercise. The Support Vector Machine (SVM) model was constructed to rank the relative importance of risk factors. Partial Least Squares (PLS) Regression was implemented as a benchmarking tool to validate the study results.

Results: Of the 12 variables analyzed for the relationship with obesity rate, 7 were statistically significant. The factors having the highest impact on the obesity rate were: hypertension, cholesterol, poverty rate, and per capita income.

Conclusion/Implications: Obesity is an underlying factor currently linked to the mortality associated with the leading causes of death in America.  The study showed that obesity is influenced by behavioral and environmental factors. Knowing the existence of these associations may be pertinent in developing interventions and public policies in the foreseeable future.

Learning Areas:

Biostatistics, economics
Public health or related research

Learning Objectives:
compare the important risk factors for obesity between SVM and PLS regression models; and demonstarte the model consistency, model fitting, and predictive accuracy

Keyword(s): Obesity, Methodology

Presenting author's disclosure statement:

Qualified on the content I am responsible for because: I am a Biostatistics professor who has taught Biostatistics to graduate students and medical residents for more than two decades.
Any relevant financial relationships? No

I agree to comply with the American Public Health Association Conflict of Interest and Commercial Support Guidelines, and to disclose to the participants any off-label or experimental uses of a commercial product or service discussed in my presentation.