Structural Properties, Cycloid Trajectories and Non-Asymptotic Guarantees of EM Algorithm for Mixed Linear Regression PubID: pubid: 0000–0000 © 2025 IEEE
Zhankun Luo, Abolfazl Hashemi · 2025
Read Structural Properties, Cycloid Trajectories and Non-Asymptotic Guarantees of EM Algorithm for Mixed Linear Regression PubID: pubid: 0000–0000 © 2025 IEEE by Zhankun Luo, Abolfazl Hashemi (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-2511-04937.txt.
Introduction
Zhankun Luo and a co-author study the expectation-maximization method for mixed linear regression with two parts, where both the mix weights and the fits are unknown. Past work covered known, balanced weights and fast rates in clean cases, but the full unknown case stayed murky.
The authors map the method's structural traits and show its path follows a cycloid-like curve. They then give non-asymptotic bounds, which state how close the method is after a set count of steps. The work fills a gap in the theory of a widely used fitting method.
Catalog record
| CREATOR | Zhankun Luo, Abolfazl Hashemi |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 265 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2511-04937