Researchers have developed a novel Web--Knowledge--Web (W$ o$K$ o$W) pipeline to address coverage gaps in business databases for domain-specific supplier discovery. This method iteratively crawls web sources, extracts structured knowledge using few-shot LLM prompting, and then uses this knowledge graph to guide further crawling toward under-represented areas. Experiments in the semiconductor equipment manufacturing sector showed improved precision and F1 scores while significantly reducing the number of pages crawled. AI
Summary written by gemini-2.5-flash-lite from 1 source. How we write summaries →
IMPACT Enhances the ability to discover niche suppliers, potentially improving supply chain resilience and market intelligence.
RANK_REASON This is a research paper describing a novel methodology for web crawling and knowledge graph construction.