Top-down extraction of semi-structured data

Abstract

We propose an innovative approach to extracting semi-structured data from Web sources. The idea is to collect a couple of example objects from the user and to use this information to extract new objects from new pages or texts. We propose a top-down strategy that extracts complex objects, decomposing them in objects less complex, until atomic objects have been extracted. Through experimentation, we demonstrate that with a small number of given examples, our strategy is able to extract most of the objects present in a Web source given as input.