Flexible Querying of Multiple Data Sources through Fuzzy Summaries

This paper deals with efficient fuzzy query processing in a distributed database context. The idea suggested is to forward the query only to the data sources that are likely to provide "good answers" to the query, which entails to define a way to assess each source by means of a "compact" representation of its content. A solution based on fuzzy summaries is outlined, and different approaches to fuzzy summaries are presented and compared. The query/summary matching aspect is tackled for each of these approaches and a general processing strategy is described.