A model of approximations for representing time-varying multidimensional data

The widespread use of new technologies for data acquisition and communications is disclosing large amounts of time-varying data that organizations wish to use for monitoring or decision support purposes. The efficient management of such large data sets depends largely on the use of appropriate data structures and access methods. This issue is an important topic of research in the spatiotemporal databases community. This paper presents a novel approach for the representation of large series of time-varying multidimensional data. It is based on a model of approximations that allows creating a hierarchical data structure by using different degrees of precision for each level. The hierarchical data structure allows representing an instance of a single series of discretely or continuously changing data as an abstract data type. The paper also shows how to use this approach to represent the movement of an object within a spatiotemporal database system.