Browsing by Subject "Longitudinal data"
Now showing 1 - 1 of 1
- Results Per Page
- Sort Options
Item Bayesian partition models for local inference in longitudinal and survival data(2021-07-29) Paulon, Giorgio; Sarkar, Abhra; Müller, Peter, 1963 August 9-; Zhou, Mingyuan; Chandrasekaran, BharathThis dissertation proposes novel Bayesian semiparametric and nonparametric methods for complex, large and potentially high-dimensional longitudinal and survival data. The first part, comprising the bulk of this thesis, develops sophisticated dynamic partition models for longitudinal data that allow common features to be shared across some time segments while differing across others. These ideas are then specifically adapted to develop novel drift-diffusion models for the analysis of behavioral data on category learning in auditory neuroscience. The second part of this work proposes a bivariate survival regression method, borrowing information across two outcomes via common features in parts of the induced marginal partitions. In terms of flexibility and interpretability, the methods presented here provide significant improvements over many previously available tools and techniques, leading to interesting, novel and meaningful inference in many diverse application areas.