Distributed semantic document retrieval using O-FCN

Object-Fuzzy Concept Network (O-FCN) is a recent knowledge representation model to integrate Fuzzy Ontologies in Information Retrieval systems. O-FCNs handle huge data collections and have to face the inherent complexity of semantic manipulation during the retrieval process. Therefore their distribution is an essential requirement to reach good scalability. We present ‘Grid2Peer’: a distributed architecture for O-FCN-based semantic information retrieval that exploits the self-organization characteristics of both Grid and P2P systems. The most relevant features in Grid2Peer are the adoption of the fuzzy sets to organize the overlay itself, the capability of migrating knowledge towards the location where it is accessed, and granting dynamic load balancing among peers. Numerical simulations are performed in order to analyze these characteristics, evaluating also fuzzy precision and fuzzy recall measures given by the distributed retrieval algorithm for the Grid2Peer architecture.