Extraction of Business Information on the Web to Supply a Geolocated Search Service

摘要

Searching information about local businesses is not a trivial problem to address. Most of existing services are supplied with manually recorded data. Based on the observation that more and more businesses are referenced on the web, we propose a new approach, which consists to extract companies' targeted information (addresses, activities, jobs, products, emails, fax, phone numbers) from websites, to supply a local business search service. The information retrieval module combines thematic, spatial and full-text criteria.