scroll
Glossary

Training Data

[ trey-ning dey-tuh ]
noun

Training data is the set of examples used to teach a machine learning model, such as messages labeled with the customer's intent or documents labeled with their type. Its quality and coverage largely determine how well the model performs.

Why it matters in regulated industries

Training data in healthcare, insurance and finance can include personal or protected information, so how it is collected, stored and used must meet privacy rules such as HIPAA and GDPR. See Ushur security and compliance.