Supporting query over dynamic combination of data sources for social media

Abstract

Meta search and social media are combined in our work to build a platform to share cross-media data. Data are extracted from web on demand with meta-search engine since more accurate and complete retrieval results can be provided than a single SE. The key to affect the performance of data extraction is the schedule of data sources to provide data efficiently. A uniform model to describe data sources is suggested, with this model a method to construct query plan graph and an algorithm to get the optimal query plan are proposed. A dynamic and adaptive adjustment approach for query plan execution is described to deal with unexpected failure of data source connection or data extraction. The details of wrapper manager encapsulating data sources are explained at the end of the paper. © 2012 Springer-Verlag.