Use the evaluate() function to evaluate on data using model rank in leaderboard and generate performance metrics. Returns DataFrame with performance metrics.
If both fit and load methods are called before predict, then fit method model will be used for prediction by default unless 'use_loaded_models' is set to True in predict.
evaluate() is not available for clustering "task_type".
Required Argument
- data
- Specifies the dataset on which performance metrics needs to be generated.Target column used for generating model is mandatory in data for evaluation.
Optional Arguments
- rank
- Specifies the rank of the model available in the leaderboard to be used for evaluation.
Default value: 1
- use_loaded_models
- Specifies whether to use loaded models from database for prediction or not.
Default value: False