Databases Reference
In-Depth Information
realistic, and subjective) to the measurement of decision support systems
information can be found.
14.3.2 Design Issues
Unfortunately, few proposals consider data quality to be a crucial factor in
the DB design process. Works like [17] and [51] are the exception in this
sense. The authors of these works provide a methodology that complements
traditional DB methodologies (e.g., [22]). At the first stage of this methodol-
ogy (see Figure 14.7), in addition to creating the conceptual schema using,
for example, an extended E/R model, we should identify quality require-
ments and candidate attributes. Thereafter, the quality parameter view
must be determined, associating a quality parameter with each conceptual
schema element (entity, relationship,
). For example, for an academic
mark, two parameters can be accuracy and timeliness. Next, subjective
parameters are objectified by the addition of tags to conceptual schema
attributes. For example, for the academic mark we can add the source of the
mark (to know its accuracy) and the date (to know its timeliness). Finally,
different quality views are integrated.
ΒΌ
Application requirements
Determine the view of data
quality requirements
quality attributes
Application view
Determine parameters
Parameter view
Determine indicators
Quality view
Quality view integration
Quality schema
Figure 14.7
Considering data quality in DB design.
