Abstract:
Current distributed IR approaches are not readily
applicable for P2P scenarios. The high dynamics in these networks
and the high cost for building and maintaining indices over
Distributed Hashtables make full text indexing and information
processing difficult to scale for large P2P networks.
My work will propose new approaches for
enabling distributed IR over P2P without limiting the network size or
mutilating the IR. The basis of these approaches is an innovative
distributed clustering algorithm, which can cluster peers in a P2P
network based on their content similarity. This clustering enables
significant network savings and enables new families of distributed
IR algorithms.
@inproceedings{papapetrou:fultextindexing,
author = {Odysseas Papapetrou},
title = {Full-text indexing and Information Retrieval in P2P Systems},
booktitle = {EDBT Phd workshop, EDBT 2008},
year = {2008}
}