Automatically extracting local ontologies via HTML tables
Xiao Tian-yuan · Computer Integrated Manufacturing Systems · 2007
Through analyzing characteristics of HTML tables in Web information sources,a method of automatically extracting local ontology via HTML tables was presented. This method consisted of four basic steps:(1) adopting two filtering rules to distinguish between position-tables and concept-tables,(2) formalizing HTML tables,(3) using statistics to decide which cells were attribute cells,(4) employing the position relationships between attribute cells and between HTML tables to ascertain the relationships among attributes.Finally,the accuracy of this method was validated by experiments.