YModel stores parameters for building a model in the format of JSON strings that saved as a file with mcf extension.
When users execute model building and batch score data containing the target variable directly on the YModel’s IDE, the model performance information will be presented in charts. When they embed the YModel into a third-party application or access YModel which is used as an external library (Read External Library Guide to learn details), the model performance information will be returned in the format of JSON strings. Developers can configure the JSON-style parameters to get personalized effects, such as visualization and user-defined charts.
Now let’s walk you through all the JSON-style parameters for model building and for getting model performance directly or based on the scoring result.