A data structure for representing aggregate data

Aggregate data are obtained by applying statistical aggregations and statistical analysis functions over raw data. We believe that a major adherence of aggregate data to the process that generates them can make their modeling simpler and more effective. We propose a formal description of the aggregation process and derive from this description a data structure for the representation of aggregate data. Several new concepts are introduced, in particular the implicit attributes, and an extension of the aggregation operator, called the generalized summarization. We also show that the introduced data structure allows us to identify in a simple way the population of a statistical object.