Score-Based Causal Discovery with Temporal Background Information
Tobias E. Larsen, Claus T. Ekstrøm, Anne H. Petersen · 2025
Read Score-Based Causal Discovery with Temporal Background Information by Tobias E. Larsen, Claus T. Ekstrøm, Anne H. Petersen (2025) 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/arxiv-2502-06232.txt.
Description
Tobias Larsen, Claus Ekstrom, and Anne Petersen wrote this statistics paper, published in two thousand twenty-five, on learning cause-and-effect maps when time gives part of the answer. If some variables are known to come earlier, whole arrows in a causal graph can be ruled in or out at once.
Their algorithm, TGES, is a score-based search that bakes tiered time knowledge into its scoring rule. They prove it always returns a graph of a stated kind, maximally oriented within its class. In the large-sample limit, that graph contains the true causal map. Simulations show gains over a constraint-based rival method, with a better trade between hits and false alarms. A worked example on life-course health data shows the method in use. The paper supplies both the tool and the terms to talk about tiered causal search.
Catalog record
| CREATOR | Tobias E. Larsen, Claus T. Ekstrøm, Anne H. Petersen |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 105 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2502-06232