< Back to Glossary

Noise

Noise is typically thought of as unexplained variability in data. Noise is in contrast to a signal, which is clearly identifiable and deterministic patterns in data. Due to the automated nature of data extraction and integration in many knowledge graphs, noise can become an issue without human intervention or self-correcting systems. 
Issues with noise within a given knowledge graph build may affect the comprehensiveness or accuracy of data coverage. And knowledge graph systems that can effectively deal with noise issues are demonstrably more reliable for business applications.