Description logic based knowledge representation for information extraction and query processing

Keyword-based search engines provide poor recall due to the fact that they ignore specialization/generalization and synonym handling. They also exhibit low precision that is retrieval with large amount of irrelevant information. This is due to the fact that the knowledge representation of the document is not well established. This work attempts to capture effectively the knowledge of the document using knowledge representation formalism Description Logic (DL) and represents the concepts as a DL hierarchy. The greatest advantage of representing knowledge in DL is that the system is now able to handle inconsistency and incompleteness present in the user queries. This is useful to the naive user who is unfamiliar in framing structured queries. Finally the performance of the proposed system is evaluated with that of content word based retrieval based on the performance metrics precision and recall.