Data Quality
The fitness of dataset(s) for an AI system’s intended purpose, judged by attributes like relevance, accuracy, representativeness, completeness and provenance.
What it triggers
- ISO/IEC 5259-1:2024 sets out that data quality encompasses dimensions such as accuracy, completeness, consistency, timeliness, and relevance which are critical for reliable AI model performance.source
- ISO/IEC 5259-1:2024 describes data quality as the degree to which data characteristics satisfy the stated and implied needs for a particular AI or machine learning application or use case.source
Definition
In the context of AI and machine learning, data quality refers to the degree to which data characteristics satisfy the stated and implied needs for a particular application or use case. It encompasses various dimensions such as accuracy, completeness, consistency, timeliness, and relevance, which are critical for reliable AI model performance.
Not to be confused with
Data provenance specifically documents the origin, collection, and transformations of data, whereas data quality evaluates broader attributes like accuracy, completeness, and relevance of the data itself.
Data governance involves policies, roles, and processes to manage data across the AI lifecycle, while data quality focuses on the fitness of the data for its intended purpose based on specific attributes.
Questions and answers
What dimensions does ISO define for data quality in AI?
ISO/IEC 5259-1:2024 defines data quality dimensions such as accuracy, completeness, consistency, timeliness, and relevance. These are critical for reliable AI model performance.
How is data quality defined for AI and machine learning use cases?
In the context of AI and machine learning, data quality refers to the degree to which data characteristics satisfy the stated and implied needs for a particular application or use case.
Why is data quality important for AI model performance?
Data quality dimensions like accuracy, completeness, consistency, timeliness, and relevance are critical for reliable AI model performance according to ISO/IEC 5259-1:2024.
Related Terms
Training Data
Data used to develop and train AI models to perform their intended functions....
Data Governance
Policies, roles and processes that ensure data used across the AI lifecycle is fit for purpose, traceable, secure and managed in line with legal and operational requirements....
Data Provenance
Documentation of data origin, collection, and transformations for AI systems....
Bias Mitigation
Measures to identify and reduce discriminatory outcomes in AI systems....