A Holdout Set is another name for “test set”, which means that it is a subset of the dataset that is held out, and not used for training algorithms. This set usually provides a final estimate of the machine learning model’s performance after training and validation.
Get early access to cnvrg.io Metacloud
Everything you need to build and deploy AI
Choose the best ML infrastructure for the job On-Demand
Leverage your entire AI ecosystem from one platform