Databases Reference
In-Depth Information
7.1
Data Quality Concepts
In this section, we introduce the basic concepts of data quality to help the readers un-
derstand these terminologies in their consequent usage.
Data Quality.
The term
data quality
is commonly conceived as a multi-dimensional
construct with a popular definition as the "fitness for use" [72]. In case of the Semantic
Web, there are varying concepts of data quality such as the semantic metadata on the
one hand and the notion of link quality on the other. There are several characteristics of
data quality that should be considered i.e. the completeness, accuracy, consistency and
validity on the one hand and the representational consistency, conciseness as well as the
timeliness, understandability, availability and verifiability on the other hand.
Data Quality Problems.
ect the potentiality of the appli-
cations that use the data are termed as data quality problems. The problems may vary
from the incompleteness of data, inconsistency in representation, invalid syntax or in-
accuracy.
A set of issues that can a
ff
Data Quality Dimensions and Metrics.
Data quality assessment involves the measure-
ment of quality
dimensions
(or
criteria
) that are relevant to the user. A data quality
assessment
metric
(or
measure
) is a procedure for measuring an information quality di-
mension [24]. The metrics are basically heuristics designed to fit a specific assessment
situation [89]. Since the dimensions are rather abstract concepts, the assessment metrics
rely on quality
indicators
that can be used for the assessment of the quality of a data
source w.r.t the criteria [47].
Data Quality Assessment Method.
A data quality assessment methodology is the pro-
cess of evaluating if a piece of data meets the information consumers need for a specific
use case [24]. The process involves measuring the quality dimensions that are relevant
to the user and comparing the assessment results with the users quality requirements.
7.2
Linked Data Quality Dimensions
In [167], a core set of 26 di
erent data quality dimensions were reported that can be
applied to assess the quality of Linked Data. These dimensions are divided into the
following groups:
ff
-
Contextual dimensions
-
Trust dimensions
-
Intrinsic dimensions
-
Accessibility dimensions
-
Representational dimensions
-
Dataset dynamicity
Figure 34 shows the classification of the dimensions into these 6 di
ff
erent groups as
well as the relations between them.
