Subgroup identification with virtual twins and GUIDE algorithms : an application to adult fitness data.

Cardinal Scholar

Show simple item record

dc.contributor.advisor Begum, Munni, 1970-
dc.contributor.author Sumy, Mst Sharmin Akter
dc.date.accessioned 2020-10-27T13:52:34Z
dc.date.available 2020-10-27T13:52:34Z
dc.date.issued 2020-07-18
dc.identifier.uri http://cardinalscholar.bsu.edu/handle/123456789/202449
dc.description Access to thesis permanently restricted to Ball State community only en_US
dc.description.abstract Cardiorespiratory fitness (CRF) is often used as an indicator of physical fitness that has an inverse association with mortality. It is practical to identify subgroups where CRF exerts positive or negative association with mortality status than on the entire population. We considered Classification and Regression tree-based algorithms: GUIDE and Virtual Twins (VT) to find subgroups of participants where CRF exerts positive or negative association with mortality status from the Ball State Adult Fitness Program Longitudinal Lifestyle Study data. Results are compared with traditional penalized Logistic regression with LASSO penalty. The results indicate that VT for classification and regression tree identified subgroups by considering BMI (Body Mass Index) and age as important predictors for different thresholds. We observed that when the threshold was larger than some specific value, VT couldn’t identify any subgroups. GUIDE selected subgroups with BMI, age, and sex as important predictors. Finally, Logistic regression with LASSO penalty found the interaction with fitness rank and other predictors: age, sex, smoking status, and hypertension. Thus, our results suggest that tree-based methods identified fewer predictors compared to the non-tree based logistic regression method. en_US
dc.description.sponsorship Department of Mathematical Sciences
dc.subject.lcsh Logistic regression analysis
dc.subject.lcsh Trees (Graph theory)
dc.subject.lcsh Cardiopulmonary fitness
dc.title Subgroup identification with virtual twins and GUIDE algorithms : an application to adult fitness data. en_US
dc.description.degree Thesis (M.S.) en_US


Files in this item

This item appears in the following Collection(s)

  • Master's Theses [5510]
    Master's theses submitted to the Graduate School by Ball State University master's degree candidates in partial fulfillment of degree requirements.

Show simple item record

Search Cardinal Scholar


Browse

My Account