Handling Failures in Data Quality Measures

Abstract

Successful data quality (DQ) measure is important
for many data consumers (or data guardians) to decide on the
acceptability of data of concerned. Nevertheless, little is known
about how “failures” of DQ measures can be handled by data
guardians in the presence of factor(s) that contributes to the
failures. This paper presents a review of failure handling mechanisms
for DQ measures. The failure factors faced by existing DQ
measures will be presented, together with the research gaps in
respect to failure handling mechanisms in DQ frameworks. In
particular, by comparing existing DQ frameworks in terms of: the
inputs used to measure DQ, the way DQ scores are computed and
they way DQ scores are stored, we identified failure factors
inherent within the frameworks. Understanding of how failures
can be handled will lead to the design of a systematic failure
handling mechanism for robust DQ measures.