What are the available configurations in an APP?
This article outlines app configurations in the AI & Analytics Engine.
During the app creation, users have the opportunity to configure the train-test split in the app.
Caution: Once the app is created, the configurations cannot be changed.
Train-test split
By default, all Apps will use a 80-20 train-test split as recommended by our Data Scientists at PI.EXCHANGE.
Users can change the train-test split ratio under the Advanced Configuration option. Simply input the desired train split amount. For example, if one desires a 70-30 train-test split, input "70".
- Train split: Used to train Machine Learning models
- Test split: Used to evaluate the performance of the trained models