Optimizing Semantic Information Retrieval by Labeling and Ontology

To optimize the semantic information retrieval by labeling and ontology methods.

This applied research has been done with the approach of content analysis. 313 Persian articles on the subject of information retrieval were collected in a database with subject search capabilities for both pre-test and post-test groups. After labeling 5700 words with the help of Ferdowsi University of Mashhad's software for natural language processing software, the ontology of concepts and their semantic relations were designed and implemented in protégé software. The accuracy of the retrieved results was measured in two stages before and after the test.

The significance level of Z test, in terms of statistical and reliability of 0.99, showed a significant difference between the accuracy of the retrieved related results in the two groups of pre-test and post-test. Therefore, these tools are acceptable.

Tow methods of natural language processing and ontology optimize semantic information retrieval.