Entity extraction from the web with WebKnox

Abstract

This paper describes a system for entity extraction from the web. The system uses three different extraction techniques which are tightly coupled with mechanisms for retrieving entity rich web pages. The main contributions of this paper are a new entity retrieval approach, a comparison of different extraction techniques and a more precise entity extraction algorithm. The presented approach allows to extract domain-independent information from the web requiring only minimal human effort. © Springer-Verlag Berlin Heidelberg 2010.