POIFormer: A Transformer-Based Framework for Accurate Point-of-Interest Attribution
Nripsuta Ani Saxena, Shang-Ling Hsu, Mehul Shetty, Omar Alkhadra, Cyrus Shahabi, Abigail L. Horn · 2025
Read POIFormer: A Transformer-Based Framework for Accurate Point-of-Interest Attribution by Nripsuta Ani Saxena, Shang-Ling Hsu, Mehul Shetty, Omar Alkhadra, Cyrus Shahabi, Abigail L. Horn (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-13070009.txt.
Description
Saxena and five co-authors matched phone visits to places in this paper from twenty twenty-five. Telling which shop or venue a person actually visited is basic to mobility analytics. GPS error and dense city venues defeat simple distance rules. Their system, POIFormer, uses a transformer to weigh many signals at once.
The model reads distance, timing, visit length, venue meaning, and crowd habits together. It also looks at the visits before and after the one in question. Crowd patterns enter through precomputed density maps, avoiding hard-to-get data layers. Tests on real mobility data beat older methods by wide margins. Gains were largest in noisy, crowded settings where the task is hardest.
Catalog record
| CREATOR | Nripsuta Ani Saxena, Shang-Ling Hsu, Mehul Shetty, Omar Alkhadra, Cyrus Shahabi, Abigail L. Horn |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 50 KB |
| VERSION | 1.0 |
| LICENSE | CC BY 4.0 |
| RIGHTS | CC BY |
| SOURCE | Europe PMC |