In the present work we suggest and test new process of web information extraction. Proposed method consider DOM tree of the web page including it's visual cues. Basic and the rst part is semantic parts extraction of a page using VIPS algorithm. Next step is validation and eventual modication of gained information based on the local context. Final part is classication of analyzing page into predened classes using got facts. Set of critics implemented by congurable instances of neural networks determine the classes.
Identifer | oai:union.ndltd.org:nusl.cz/oai:invenio.nusl.cz:300188 |
Date | January 2010 |
Creators | Pekárek, Filip |
Contributors | Galamboš, Leo, Kopecký, Michal |
Source Sets | Czech ETDs |
Language | Czech |
Detected Language | English |
Type | info:eu-repo/semantics/masterThesis |
Rights | info:eu-repo/semantics/restrictedAccess |
Page generated in 0.0015 seconds