Detail publikace

Information Extraction from Web Sources based on Multi-aspect Content Analysis

Originální název

Information Extraction from Web Sources based on Multi-aspect Content Analysis

Anglický název

Information Extraction from Web Sources based on Multi-aspect Content Analysis

Jazyk

en

Originální abstrakt

Information extraction from web pages is often recognized as a  diffcult task mainly due to the loose structure and insuffcient semantic annotation of their HTML code. Since the web pages are primarily created for being viewed by human readers, their authors usually do not pay much attention to the structure and even validity of the HTML code itself. The CEUR Workshop Proceedings pages are a good illustration of this. Their code varies from an invalid HTML markup to fully valid and semantically annotated documents while preserving a kind of uni ed visual presentation of the contents. In this paper, as a contribution to the ESWC 2015 Semantic Publishing Challenge, we present an information extraction approach based on analyzing the rendered pages rather than their code. The documents are represented by an RDF-based model that allows to combine the results of difffferent page analysis methods such as layout analysis and the visual and textual feature classi cation. This allows to specify a set of generic rules for extracting a particular information from the page independently on its code.

Anglický abstrakt

Information extraction from web pages is often recognized as a  diffcult task mainly due to the loose structure and insuffcient semantic annotation of their HTML code. Since the web pages are primarily created for being viewed by human readers, their authors usually do not pay much attention to the structure and even validity of the HTML code itself. The CEUR Workshop Proceedings pages are a good illustration of this. Their code varies from an invalid HTML markup to fully valid and semantically annotated documents while preserving a kind of uni ed visual presentation of the contents. In this paper, as a contribution to the ESWC 2015 Semantic Publishing Challenge, we present an information extraction approach based on analyzing the rendered pages rather than their code. The documents are represented by an RDF-based model that allows to combine the results of difffferent page analysis methods such as layout analysis and the visual and textual feature classi cation. This allows to specify a set of generic rules for extracting a particular information from the page independently on its code.

BibTex


@inproceedings{BUT119821,
  author="Martin {Milička} and Radek {Burget}",
  title="Information Extraction from Web Sources based on Multi-aspect Content Analysis",
  annote="Information extraction from web pages is often recognized as a  diffcult task
mainly due to the loose structure and insuffcient semantic annotation of their
HTML code. Since the web pages are primarily created for being viewed by human
readers, their authors usually do not pay much attention to the structure and
even validity of the HTML code itself. The CEUR Workshop Proceedings pages are
a good illustration of this. Their code varies from an invalid HTML markup to
fully valid and semantically annotated documents while preserving a kind of
unied visual presentation of the contents. In this paper, as a contribution to
the ESWC 2015 Semantic Publishing Challenge, we present an information extraction
approach based on analyzing the rendered pages rather than their code. The
documents are represented by an RDF-based model that allows to combine the
results of difffferent page analysis methods such as layout analysis and the
visual and textual feature classication. This allows to specify a set of generic
rules for extracting a particular information from the page independently on its
code.",
  address="Springer International Publishing",
  booktitle="Semantic Web Evaluation Challenges, SemWebEval 2015 at ESWC 2015",
  chapter="119821",
  doi="10.1007/978-3-319-25518-7_7",
  edition="Communications in Computer and Information Science",
  howpublished="electronic, physical medium",
  institution="Springer International Publishing",
  number="548",
  year="2015",
  month="may",
  pages="81--92",
  publisher="Springer International Publishing",
  type="conference paper"
}