Keeping track of the evolution of medical records is a critical part of a clinical data warehouse. Time has an important role in the clinical domain to help understand and analysis different kinds of clinical procedures such as diagnosis, treatment, and prevention. More details about the importance of time in the clinical domain can be found in this article [Madkour et al. 2016]. Temporal clinical data warehouses are acquiring increasing importance in the health field in order to mine massive data and therefore for constructing the clinical timeline [Madkour et al. 2016]. Information variation over time is crucial for most analysis purposes. Having a well-defined temporal schema ensures correct temporal semantic and temporal constraint management. The DW schema must be based on a sound, comprehensive and formalized temporal model to improve expressiveness and interoperability. A sound temporal schema plays an essential role in minimizing data uncertainty, data indeterminacy and query expressiveness. Current temporal data models [Date et al. 2014; Johnston and Weis 2010; Snodgrass 2000] relies on relational data model to define design guidelines and constraints regarding temporal representations and integrity. Some methods rely on ad hoc models that might work with requirement-driven methods but carry limitations. In fact, when applied to a context where prospective operations are not pre-defined, it becomes essential to have a temporal model which stands on its own, provides intrinsic computability soundness, and gives (automatable) provable transformation rules. This technical report demonstrates the integration and generalization of two major temporal models (the Bitemporal Conceptual Data Model and the Date-Darwen-Lorentzos Model) in terms of views defined using the Unified Bitemporal Historicization Framework. Using this framework, historicization is defined as a suite of simple automated steps. The primary aim of this work is to help database designers to model historicized schema based on a sound theory ensuring a sound temporal semantic, data integrity, query expressiveness and guided automation.
Report : UBHF_Technical_Report.pdf
Appendix : UBHF_Appendix.pdf