Clustering longitudinal multivariate trajectories using an ensemble of principal component trees
Bastian Pfeifer, Simon Grabner, Andrea Berghold, Markus Loecher · 2026
Read Clustering longitudinal multivariate trajectories using an ensemble of principal component trees by Bastian Pfeifer, Simon Grabner, Andrea Berghold, Markus Loecher (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/pmc-13474914.txt.
Description
Pfeifer and colleagues present a method in 2026 for grouping patients or samples by how their measurements change over time, a task called longitudinal clustering. Repeated measures are often messy, high-dimensional, and taken at irregular times, which breaks simpler grouping methods. Their framework, TAPIO, repeatedly samples small feature subsets, compresses them with principal components, a way of summarizing correlated measures, and merges many tree-based groupings into one stable result.
Three variants handle different data shapes, covering tidy repeated measures, single samples over time, and irregularly timed multivariate records in the method family described. In simulations the approach recovered the true groups and flagged the features that defined them, matching or beating rival methods on the tests run. Applied to a heart patient cohort it found subgroups with distinct survival patterns, and on protein data over time it separated molecularly distinct clusters with readable protein importance profiles.
Catalog record
| CREATOR | Bastian Pfeifer, Simon Grabner, Andrea Berghold, Markus Loecher |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 129 KB |
| VERSION | 1.0 |
| LICENSE | CC BY 4.0 |
| RIGHTS | CC BY |
| SOURCE | Europe PMC |