Jessica Trail, Ph.D.

This graduate-level course covers methodology and software used in the analysis and exploration of large complex data sets. Topics include statistical learning; resampling methods; linear regression; variable selection; regression shrinkage; dimension reduction; non-linear methods; logistic regression, discriminant analysis; nearest-neighbors; decision trees; bagging; boosting; support vector machines; principal components analysis; clustering.

I was awarded a pre-doctoral fellowship grant from the National Institutes for Health, and the Nationl Institute on Drug Abuse. The title of the study is "New Dynamic Models for Intensive Longitudinal Substance Use Data" [NIH/NIDA F31DA032202 (PI: Trail)]. The purpose of this grant was to develop new statistical methods for the analysis of intensive longitudinal data (ILD), and my dissertation research involves developing new methods and applications for dynamic models with time-varying effects. My specific aims are (1) to use penalized splines to model process dynamics in ILD, (2) to develop improved methods for testing hypotheses about time-varying effects, and (3) to demonstrate the effectiveness of the proposed methods by applying them to a rich, longitudinal data set on daily substance use among college freshmen. The new hypothesis testing methods will enable researchers to test hypotheses about whether an effect significantly varies over time, as well as whether or not a time-varying effect is significant.

As part of this position I integrated techniques from functional data analysis and dynamical systems models for system identification of behavioral processes.

Instructor for an introductory statistics course