Stat Cafe - Gayun Kwon
Functional Principal Component Analysis for Network-Dependent Data
- Time: Tuesday, 10/27/2026 from 11:10AM to 12:25PM
- Location: BLOC 448
Description
Functional data analysis often assumes independent and identically distributed curves, yet curves can be dependent. Ecological momentary assessment (EMA) adds another challenge by collecting ordinal, binary, truncated, and continuous non-Gaussian responses. We propose a unified functional principal component analysis framework that represents dependence between curves through a graph and accommodates these different data types. Using EMA data from the National Institute of Mental Health Family Study of Mood Spectrum Disorders, we illustrate how recovering latent curves and functional principal component scores helps characterize individuals’ typical daily mood patterns and day-to-day deviations from those patterns.
Our Speaker
Gayun Kwon is a Ph.D. student in the Department of Statistics at Texas A&M University, co-advised by Dr. Debangan Dey and Dr. Nilanjana Laha. Her research interests include mixed-type functional data analysis and its applications to mental health and ecological momentary assessment (EMA). She earned her B.S. and M.S. in Statistics from Ewha Womans University in South Korea.