EvoTaxo: Building and Evolving Taxonomy from Social Media Streams
Constructing taxonomies from social media corpora is challenging because posts are short, noisy, semantically entangled, and temporally dynamic. Existing taxonomy induction methods are largely designed for static corpora and often struggle to balance robustness, scalability, and sensitivity to evolving discourse. We propose EvoTaxo, a LLM-based framework for building and evolving taxonomies from temporally ordered social media streams. Rather than clustering raw posts directly, EvoTaxo converts each post into a structured draft action over the current taxonomy, accumulates structural evidence over time windows, and consolidates candidate edits through dual-view clustering that combines se...
| Authors | Yiyang Li, Tianyi Ma, Yanfang Ye |
|---|---|
| Primary category | cs.CL |
| Published | 20 March 2026 |
| Last updated at source | 20 March 2026 |
Silytics shows public metadata from arXiv and links to the original. Information may be incomplete or out of date. Last refreshed 17 October 2026. Licenses and terms of the original apply to the work itself. Not financial, legal or professional advice.