Nowadays searching over structure P2P network mainly use literal matching words occurrence in documents with those present in user queries. But this process of searching can not deal with the semantics and semantic relationships of words. In this paper, we address these problems by proposing semantic search over structure P2P network based on ontology. We do not assume a global ontology but heterogeneous peer ontologies; instead, each peer maintains an own ontology and constructs concept descriptions for documents and query based on it. Two levels of distributed index are constructed: distributed concept index for peer ontologies and distributed inverted index by concept for documents. Based on them, semantic search is implemented with inference technique of Description Logic between concept descriptions of documents and query. Simulation experiments show that our approach is much higher on retrieval perform than traditional method of literal keywords matching.
展开▼
