engleski

A Protocol For Self-Organizing Peer-to-Peer Network Supporting Content-Based Search

For a peer-to-peer(P2P) content sharing network holding large amount of data, an efficient semantic based search mechanism is a key requisite. Semantic based search should generate as little traffic (messages) possible while achieving precision and recall rates comparable to those of correspondent centralized system. In this paper protocols for self-organizing P2P networks that arranges links between peers according to peer's content are developed and tested. Peers organize themselves into "semantic communities" without losing links to other semantic communities. Proposed network requires no prior knowledge of the semantics of documents that are to be shared in the system. Through simulations, it is shown that proposed network is resilient to membership changes and achieves high recall rates.

Peer-to-peer; Content-based search; Information retrieval; Algorithm; Semantic

nije evidentirano