%0 Unpublished Work %1 anton-lwa05 %A Anton, Tobias %D 2005 %I GI-Fachgruppen ABIS, AKKD, FGML %J LWA 2005 - Workshopwoche der GI-Fachgruppen/Arbeitskreise %K induction, XML, Wrapper DOM XPath, machine space learning, version %P 126-133 %T XPath-Wrapper induction by generalizing tree traversal patterns %U http://anton.dnsalias.org/public/tobias/publications/anton-lwa05.pdf %X We introduce a wrapper induction algorithm for extracting information from tree-structured docu-ments like HTML or XML. It derives XPath-compatible extraction rules from a set of anno-tated example documents. The approach builds a minimally generalized tree traversal pattern, and augments it with conditions. Another variant se-lects a subset of conditions so that (a) the pattern is consistent with the training data, (b) the pat-tern’s document coverage is minimized, and (c) conditions that match structures preceding the target nodes are preferred. We discuss the ro-bustness of rules induced by this selection strat-egy and we illustrate how these rules exhibit knowledge of the target concept.

@unpublished{anton-lwa05, abstract = {We introduce a wrapper induction algorithm for extracting information from tree-structured docu-ments like HTML or XML. It derives XPath-compatible extraction rules from a set of anno-tated example documents. The approach builds a minimally generalized tree traversal pattern, and augments it with conditions. Another variant se-lects a subset of conditions so that (a) the pattern is consistent with the training data, (b) the pat-tern’s document coverage is minimized, and (c) conditions that match structures preceding the target nodes are preferred. We discuss the ro-bustness of rules induced by this selection strat-egy and we illustrate how these rules exhibit knowledge of the target concept.}, added-at = {2006-07-07T16:15:42.000+0200}, author = {Anton, Tobias}, biburl = {https://www.bibsonomy.org/bibtex/2e330c0c171e8a33cfcf4c50610d3b955/tobias}, description = {A version space algorithm is used to learn extraction rules on DOM trees. The learner is biased to minimize document coverage to provide for robust extraction rules. Finally, it is shown how induced rules can be converted into statements of the XPath language.}, interhash = {0a2eb7696683298d226c942ff7ae71b3}, intrahash = {e330c0c171e8a33cfcf4c50610d3b955}, journal = {LWA 2005 - Workshopwoche der GI-Fachgruppen/Arbeitskreise}, keywords = {induction, XML, Wrapper DOM XPath, machine space learning, version}, pages = {126-133}, publisher = {GI-Fachgruppen ABIS, AKKD, FGML}, timestamp = {2006-07-07T16:15:42.000+0200}, title = {XPath-Wrapper induction by generalizing tree traversal patterns}, type = {research report}, url = {http://anton.dnsalias.org/public/tobias/publications/anton-lwa05.pdf}, year = 2005 }