Metrics for the evaluation of data quality of signal data in industrial processes

For Industry 4.0 technology as well as Cyber-Physical Production Systems analysis of data gained more and more importance. But the disparity of data, often make the efficient use of data mining methods difficult due to data with poor quality. To evaluate data quality and further adopt appropriate measures, the proposal develops a data quality model fitted to the specific properties of signal data of industrial processes. Relevant data quality characteristics are identified and a classification of these characteristics is conducted to ascertain important factors. Furthermore, a measurement for the characteristic Completeness, aggregated of its sub-dimensions, is defined. The data quality model is applied to two different use cases showing its effectiveness and validity of the defined measures. The efficient use of real industrial signal data e.g. appropriateness of the data for the specific data mining purpose, is supported by a comprehensive measurement for data quality and the detailed discussion of the influencing factors.