Graph-Based Navigation Strategies for Heterogeneous Spatial Data Sets

Querying heterogeneous spatial databases involves not only characterizing and comparing the information content of several databases, but also navigating or accessing the data sets with the query answer. This work proposes a formalism that relates the information content of data sets by three basic types of correspondence relations: data equivalence, difference of data omission, and difference of data commission. These correspondence relations define the information space over which a navigation process is carried out. Based on a complete or an incomplete information space, this work proposes strategies that optimize the retrieval process of information coming from different databases. The results of this study show the advantages of defining the information space to select and access databases. In particular, strategies that estimate the information contribution of data sets based on correspondence relations outperform a strategy that considers a random list or a list of data sets sorted by size.