DATA EXTRACTION FROM REPOSITORIES ON THE WEB: A SEMI-AUTOMATIC APPROACH

Article type: Research Article

Authors: Bayrak, Coşkun | Kolukísaoğlu, Hayrettin | Sieloff, Steve

Affiliations: Computer Science Department, University of Arkansas at Little Rock, Little Rock, AR, U.S.A. | Acxiom Corporation, Little Rock, AR, U.S.A.

Abstract: The World Wide Web (WWW) is becoming the most important source of information for business intelligence and information dissemination. Past information gathering techniques like surfing and sifting are proving insufficient in processing the vast volumes of data readily available from the Web. In addition, companies are being forced to integrate this vast data repository within specific cost, time, and reliability spectrums. This paper presents the fundamentals of a system called "Browser Harness" (B2H) that extracts the requested data from Web sites in a supervised fashion. The algorithmic background of this system is based on the tag structure of web pages, as HTML is the predominate choice for rendering web page content on the WWW. B2H is an interactive tool for harnessing data from semi-structured and structured web pages by analyzing the tag structure of the input page and locating the data in the HTML code. The extracted data is then exported to XML, delimited text, or database tables.

Keywords: Data extraction, web mining

Journal: Journal of Integrated Design & Process Science, vol. 7, no. 4, pp. 13-23, 2003