Techniques of abstraction and use of metadata structures

Metadata integration and use represents an open problem due to the rapid expansion of data resources in a nonstandard compliant environment. The increasing need to access and retrieve relevant data based on semantic similarity from distributed, autonomous and heterogeneous sources demands innovative solutions that can offer an integrated global view over various local schemas. The framework proposed in this paper aims to achieve metadata integration and aggregation while preserving initial semantics and minimal data loss. Main problems are metadata reduction to a generic structure as well as identifying only those resources that match the crawling profile.