Discovering Knowledge from Relational Data Extracted from Business News
15 Pages Posted: 13 Oct 2008
Date Written: 2002
Thousands of business news stories (including press releases, earningsreports, general business news, etc.) are released each day. Recently, informationtechnology advances have partially automated the processing ofdocuments, reducing the amount of text that must be read. Current techniques(e.g., text classification and information extraction) for full-text analysis for themost part are limited to discovering information that can be found in singledocuments. Often, however, important information does not reside in a singledocument, but in the relationships between information distributed over multipledocuments. This paper reports on an investigation into whether knowledgecan be discovered automatically from relational data extracted from large corporaof business news stories. We use a combination of information extraction,network analysis, and statistical techniques. We show that relationally interlinkedpatterns distributed over multiple documents can indeed be extracted,and (specifically) that knowledge about companiesÃÂÃÂÃÂÃÂ¢ÃÂÃÂÃÂÃÂÃÂÃÂÃÂÃÂ interrelationships can bediscovered. We evaluate the extracted relationships in several ways: we give abroad visualization of related companies, showing intuitive industry clusters;we use network analysis to ask who are the central players, and finally, weshow that the extracted interrelationships can be used for important tasks, suchas for classifying companies by industry membership.
Suggested Citation: Suggested Citation