Temporal aggregation on user-defined granularities

Abstract

Time-varying data play a major role in many applications, and, starting from the 80's, they have been widely studied in temporal databases. In the last two decades, several researchers have shown that, to deal with many application domains, user-def ined temporal granularities must be coped with. When data are stored at multiple user-defined temporal granularities, the task of defining proper conversion functions to aggregate data from an origin granularity (e.g., business days) to a task granularity (e.g., months) is of primary importance. However, current temporal database approaches mostly demand such a task to system administrators, or to specific applications, providing no methodology or general guideline to accomplish it. In this paper, we propose a general and application-independent methodology which, on the basis of the temporal relationship between two user-defined granularities, provides users with a set of conversion/aggregation functions between them, consistent with the telic vs. atelic character of the data to be aggregated. The correctness of the approach is also proved.

Temporal Data Update Methodologies for Data Warehousing

“…These applications may need a different time-slice of data. The data warehouses must cope with the temporal granularities of data (Terenziani 2012).…”

A 1NF temporal relational model and algebra coping with valid-time temporal indeterminacy

“…Notably, our group has a tradition in the application of AI techniques to cope with complex temporal phenomena in relational databases (Terenziani 2012;Anselma et al 2013aAnselma et al , 2013bTerenziani 2003;Terenziani and Snodgrass 2004;Terenziani 2013). …”