From: Generating metadata from web documents: a systematic approach
Approach | Discussion and comparisons |
---|---|
The search space depends on the acquirable schema or other semantic information. No discussions for generating semantic information from semi-structured documents. | |
The middleware or translator can map or convert schema between structured data sources and Semantic Web schema. It is infeasible on the cases of semi-structured or even un-structured data resources. | |
Schema generation from document based on knowledge engineering [7, 8] | Current schemes can generate either linguistic or semantic annotation of data pieces in web documents using prior-knowledge or NLP technologies. It is not suitable for “modeling” the web document sets or other textbases. For problem-solving or topic search purposes, the solutions are not sufficient. |
Schema generation from document based on structural part of document [21] | The structure-based approaches generate RDF only based on structural part of document. Such solution is simple to implement, while the generated RDF might not be helpful for users for question solving purposes. |