Skip to content

Independent standards discovery and compliance platform

How we work

ISO/IEC 5259-2:2024 - Data Quality Measures for ML

Part 2 focuses on measures of data quality in analytics and machine learning. It is relevant when project teams need to explain how they judge a dataset’s suitability and how the reported measure relates to a particular purpose, such as training, evaluation or operational use.

What you are buying

Before comparing data quality results, agree on the dataset boundaries and intended use. Check what was measured, how records were selected and whether a result from one sample or period still represents the data currently being used.

Frequently asked questions

Does Part 2 tell us what data quality score makes a dataset acceptable?

It describes measures you can use to assess data quality. Set acceptance thresholds for your intended use and the consequences of relying on data that falls short.

Which information should I bring to a dataset review using Part 2?

Bring the intended analysis, a description of the dataset and the decisions the results will inform. That makes it easier to choose measures suited to the actual quality problem.