HLR-UFS: Hessian-based Unsupervised Feature Selection using Low-Rank Approximation
Jiyan Zhang, Weihan Lin, Yanfang Liu, Shuzhen Tu, Ming Peng · 2026
Read HLR-UFS: Hessian-based Unsupervised Feature Selection using Low-Rank Approximation by Jiyan Zhang, Weihan Lin, Yanfang Liu, Shuzhen Tu, Ming Peng (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/plos-pone-0356837.txt.
Description
Jiyan Zhang and co-authors build a math method for picking the most useful features from data when no labels exist to guide the choice. Feature picking cuts noise and cost in machine learning, but common tools can flatten the hidden shape of a dataset. Their method uses a curvature-aware penalty to keep complex shapes intact, a compression step to drop repeated features, and a filter that resists stray outliers.
Across nine standard test datasets, the method chose feature sets that performed better than rival approaches, and the team proves their solving routine settles properly. Zhang and colleagues aim the tool at tasks such as image and gene analysis. Those datasets are wide, messy, and unlabeled. Keeping true structure matters more than raw size.
Catalog record
| CREATOR | Jiyan Zhang, Weihan Lin, Yanfang Liu, Shuzhen Tu, Ming Peng |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 59495 |
| LICENSE | CC_BY |
| RIGHTS | CC BY |
| SOURCE | PLOS |