Nonparametric Quantile Curves of Health Risk Factors for American Adolescents

• Activity Number: 585 - Recent Advances in Quantile Regression
• Type: Contributed
• Date/Time: Wednesday, August 1, 2018 : 2:00 PM to 3:50 PM
• Sponsor: Section on Nonparametric Statistics
• Author(s): Jessica Rudd* and Mohammed Chowdhury
• Companies: Kennesaw State University and Kennesaw State University
• Keywords: local polynomial; kernel smoothing; spline smoothing; bandwidth; cross validation; two-step smoothing

Abstract

Nonparametric regression plays a significant role in smoothing estimation of functional data. In this paper, we will consider several nonparametric methods for smoothing estimation of the quantile curves of health risk factors such as SBP, DBP, HDL, LDL, TC, TG, Weight and BMI for African American and Caucasian adolescents. We use a two-step smoothing procedure in which we obtain the raw estimate of quantiles at each times point (age) and then smooth these raw quantiles by available smoothing methods such as kernel smoothing, local polynomial smoothing, and spline smoothing. Application of the methods are demonstrated on a longitudinal study, NGHS(National Growth and Health Study). From the application, we see that African American adolescents have higher quantile curves in all health risk factors except TG (Triglyceride) than their Caucasian counterparts.