A hybrid approach towards wrapper induction
Stefan Raeymaekers, Maurice Bruynooghe · Lirias · 2007
Abstract. The approaches to learn wrappers for extraction from semi-structured documents (like HTML documents) are divided into string based ones, and tree based ones. In previous papers we have shown that tree based approaches perform much better and need less examples than string based approaches, but have the disadvantage that they can only extract complete text nodes, whereas string based approaches can extract within text nodes. In this paper we propose a hybrid approach that combines the advantages of both systems. We compare this approach experimentally with a string based approach on some sub node extraction tasks. 1