Computational Math and Statistics Seminar by Kyunghee Han: Shape-Constrained Inference in Function-on-Scalar Regression Models

Speaker: Kyunghee Han, assistant professor of mathematics, University of Illinois Chicago

Title: Shape-Constrained Inference in Function-on-Scalar Regression Models

Abstract:

As data size and complexity grow, functional data analysis (FDA) has become increasingly prevalent in real-world applications and a crucial framework in modern data science. Unlike traditional analytic methods that focus on statistical associations within Euclidean spaces, FDA provides essential tools for modeling associations in non-Euclidean spaces, making statistical inference within the space of functions more appropriate. In this talk, we will overview statistical inference for function-on-scalar regression models, which are widely used in various fields to explore associations between functional responses and scalar covariates. Specifically, we will focus on shape-constrained inference, which is particularly valuable for domain experts as it facilitates meaningful interpretation in practice. The utility of shape-constrained inference will be demonstrated through data examples, including clinical trials of NeuroBloc for type A-resistant cervical dystonia and the National Institute of Mental Health Schizophrenia Study.

Computational Mathematics and Statistics Seminar