MLOps Glossary

A/B Testing

A/B Testing:

Running 2 (or more) models in parallel with traffic splitting equally whilst collecting data on model performance to compare which model is best. This is traditionally applied in areas such as online advertising. Note that traditional A/B testing may not include any underlying ML models per se, but rather refers to different variants of some product, e.g. different views of a website to test which increases user retention.

SEE HOW SELDON WORKS FOR YOU

Serve, monitor, explain, and manage your models today.

© 2023 Seldon Technologies. All Rights Reserved.

Rise London
41 Luke Street
Shoreditch
EC2A 4DP

UK: +44 (20) 7193-6752
US. +1 (646) 397-9911

Email: hello@seldon.io