%0 Journal Article
%T Application of Binary Logistic Regression Model to Assess the Likelihood of Overweight
%A Noora Shrestha
%J American Journal of Theoretical and Applied Statistics
%P 18-25
%@ 2326-9006
%D 2019
%R 10.11648/j.ajtas.20190801.13
%X Abstract: This study attempts to assess the likelihood of overweight and associated factors among the young students by
analyzing their physical measurements and physical activity index. This paper has classified four hundred and fifteen subjects
and precisely estimated the likelihood of outcome overweight by combining body mass index and CUN-BAE calculated.
Multicollinearity is tested with multiple regression analysis. Box-Tidwell Test is used to check the linearity of the continuous
independent variables and their logit (log odds). The binary regression analysis was executed to determine the influences of
gender, physical activity index, and physical measurements on the likelihood that the subjects fall in overweight category. The
sensitivity and specificity described by the model are 55.9% and 96.9% respectively. The increase in the value of waist to
height ratio and neck circumference and drop in physical activity index are associated with the increased likelihood of subjects
falling to overweight group. The prevalence of overweight is higher (27.8%) in female than in male (14.7%) subjects. The odds
ratio for gender reveals that the likelihood of subjects falling to overweight category is 2.6 times higher in female compared to
male subjects.
Keywords: Overweight, Waist to Height Ratio, Neck Circumference, Binary Logistic Model, Odds Ratio
%K Overweight
%K Waist to Height Ratio
%K Neck Circumference
%K Binary Logistic Model
%K Odds Ratio
%U http://www.sciencepublishinggroup.com/journal/paperinfo?journalid=146&doi=10.11648/j.ajtas.20190801.13