Item

Abstract

One of the most challenging problems in peer-to-peer networks is \textit{query
routing}: effectively and efficiently
identifying peers that can return high-quality local results for a given query.
Existing methods from the areas of
distributed information retrieval and metasearch engines do not adequately
address the peculiarities
of a peer-to-peer network.

The main contributions of this thesis are as follows:

\begin{enumerate}
\item{Methods for query routing that take into account the mutual overlap of
different peers' collections,}
\item{Methods for query routing that take into account the correlations between
multiple terms,}
\item{Comparative evaluation of different query routing methods.}
\end{enumerate}

Our experiments confirm the superiority of our novel query routing methods over
the prior state-of-the-art,
in particular in the context of peer-to-peer Web search.