The Model Leaderboard is a feature of the AI & Analytics Engine that enables users to view a summary of their model's performances.
The AI and Analytics Engine enables users to train multiple models. Users can easily try various ML algorithms and feature sets.
Tip: For more information on the algorithms provided by the Engine, see this article. To learn more about feature sets, see this article.
When multiple models are trained, it is useful for users to get an overview of all their performances. This is the purpose of the Model Leaderboard. The Model Leaderboard shows a summary of how different models performed and ranks these models based on their performance. The Model Leaderboard presents model rankings based on prediction quality, prediction time, and training time. The user can use this to compare the models and decide which one they want to deploy.
Tip: To learn more about how to use the model leaderboard, see this article. If you are curious about how prediction quality is calculated, see this article.