Measuring Statistical Dependence via Characteristic Function IPM
Povilas Daniušis, Shubham Juneja, Lukas Kuzma, Virginijus Marcinkevičius, Boris Ryabko · 2025
Read Measuring Statistical Dependence via Characteristic Function IPM by Povilas Daniušis, Shubham Juneja, Lukas Kuzma, Virginijus Marcinkevičius, Boris Ryabko (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/pmc-12732038.txt.
Description
Daniušis and co-authors proposed a new measure of statistical dependence in a 2025 paper. Their measure works in the frequency domain and compares a joint pattern with what independence would give. It is built to stay stable under common changes and to behave well when data are reduced.
An algorithm with a warm-up step estimates the measure, and that warm-up proved key to stable results. The estimator can plug into modern learning pipelines. In tests, it spotted sparse and curved patterns that other measures missed in some cases. Across many data sets for feature picking, it beat rivals in some comparisons and lost in fewer.
Catalog record
| CREATOR | Povilas Daniušis, Shubham Juneja, Lukas Kuzma, Virginijus Marcinkevičius, Boris Ryabko |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 66 KB |
| VERSION | 1.0 |
| LICENSE | CC BY 4.0 |
| RIGHTS | CC BY |
| SOURCE | Europe PMC |