Abstract:

Indexing becomes an active area of current researches in search engines. In order to relieve the problem of semantic confusion or divergence caused by synonym and polysemy, based on the existing researches, a semantic indexing method is proposed. Through refining the contextual information with ontology in the index, the search range can be controlled well. The relevance is ensured and the missing degree is reduced by applying the concept matching retrieval approach. Additionally, a ranking method is given. The proposed method obviously improves the accuracy. The experiments show that the method enhances the precision and recall and the ranking method has a good effect.

Key words: context;semantic indexing;ontology;concept matching