How to Measure Data Quality? - A Metric-Based Approach
How to Measure Data Quality? - A Metric-Based Approach
复制标题
如何衡量数据质量？
DOI：
--
复制
发表时间：
2007
期刊：
影响因子：
--
通讯作者：
Bernd Heinrich
中科院分区：
文献类型：
--
作者：
M. Kaiser;Mathias Klier;Bernd Heinrich
数据质量的日益增长的相关性揭示了对适当测量的需要，因为量化数据质量对于以经济的方式规划质量测量是必不可少的。本文分析了数据质量如何可以量化的特定维度。首先，几个要求（例如规范化，可解释性）设计适当的度量。其次，我们分析了文献中的指标，并讨论了他们的要求。第三，在现有方法的基础上，设计了新的维度正确性和及时性度量，以满足定义的要求。最后，我们评估我们的及时性指标的案例研究：在与德国的一个主要的移动的服务提供商的合作，该方法被应用于活动管理，以提高成功率和利润。
The growing rel evance of data quality has revealed the need for adequate measurement since quantifying data quality is esse ntial for pla nning quality measures in an economic manner. This paper analyzes how data quality can be quantified with respect to particular dimensions. Firstly, several requirements are stated (e.g. normal ization, interpretability) for designing adequate metrics. Seco ndly, we analyze metrics in literature and discuss them with regard to the requirements. Thirdly, based on existing approaches new metrics for the dimensions correctness and timeliness that meet the defined requirements are designed. Finally, we evaluate o ur metric for timeliness in a case study: In cooperation with a major German mobile services provider, the approach was applied in campaign management to improv e both success rates and profits.