Empirical Study on Usefulness of Algorithm SACwRApper for Reputation Extraction from the WWW

摘要

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. 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.