A0159
Title: Advanced longitudinal functional data methods for emerging repeated measurements
Authors: Ana-Maria Staicu - North Carolina State University (United States) [presenting]
Abstract: Many longitudinal studies collect data in the form of functions or surfaces that are repeatedly observed over time. Parsimonious modeling frameworks are first introduced for such functional data that enable the extraction of low-dimensional features while accounting for the longitudinal study design. These methods are specifically developed to capture the dynamic behavior of the underlying process and are computationally efficient. The framework is then extended to accommodate pointwise skewness in the data. Finally, it is demonstrated how these techniques can be used to construct likelihood-based inference procedures for testing whether the mean function changes over time.
