Empirical study on usefulness of algorithm SACwRApper for reputation extraction from the WWW

Abstract

We consider the problem of extracting texts related to a given keyword from Web pages collected by a search engine. Recently, we proposed a method using both structural and content information[1, 2]. In our previous paper, we reported good extraction performance of our method only for Ramen-shop dataset written in Japanese. In this paper, we examine it for datasets of other kind of restaurants, and also for a dataset written in English. We discuss some modification for performance improvement. © Springer-Verlag Berlin Heidelberg 2005.