Web information extraction for content augmentation

Abstract

Today, users have to cope with an overwhelming number of TV channels and Web content sources. We introduce automatic content augmentation as a novel approach to contextual information extraction on behalf of the user where the context is provided by the primary content source (i.e. TV channel) and tailored by the user's preferences. A key aspect of this approach is Web information extraction (WebIE) which automatically derives structured information from unstructured Web documents. Our system executes WebIE tasks, each an instantiation of WebIE rules, our generic document processors. We present two WebIE approaches: diffusion WebIE that crawls a wide set of Web pages and extracts information from a subset of the pertinent pages; and laser WebIE that accesses a select set of Web pages and extracts narrowly defined information. We describe the architecture and the implementation details of the system and provide detailed laser WebIE examples.