P2P Semantic Search

We consider P2P Semantic Search as a process of finding documents, which are semantically, i.e., with respect to the meaning, related to the user information needs, in a document collections distributed among a group of peers, i.e., autonomous information sources. To organize documents stored on a single peer efficiently for the search, documents are classified to the user-generated classifications. Nodes in the classification specify concepts which the user is interested in. Accordingly, the whole classification specifies the user interest profile. To provide effective search in the P2P network, peers in the network should have some ways for cooperation. In our approach, related nodes in classifications on different peers are interconnected by means of semantic links, which allow peers in the network to reason about the contents of each other and efficiently cooperate. The main foci of the current PhD thesis are the development of an algorithm for P2P Semantic Search in a distributed system of interconnected classifications; the development of a P2P Semantic search system implementing the algorithm; and the development of a testing methodology allowing for a comprehensive evaluation of the system.