Distributed Data Mining: Server Selection and Query Processing based on Communication Costs

摘要

Distributed data mining intends to get the global knowledge from the local data at distributed site. In distributed data bases data are distributed among several sites using 1) Horizontally Fragmented Distributed Data method 2) Vertically Fragmented Distributed Data method. When the Query is posed by the user result will be evaluated at one site, as a part of this data has to be moved from local sites to global site (central site). So this paper proposes a methodology for server (central site) selection where the query has to be executed by considering the communication costs among the sites and also query processing at central site.