PublicDatasetModelMetrics
- algorithmType: stringrequired
Human-readable name of the selected algorithm.
- correlationType: string
_validation _name requiredName of the primary performance metric, such as r2_cv_mean or mcc_cv_mean.
- correlationType: number
_validation _value requiredPrimary cross-validation or test-set performance value. Interpret it using correlation_validation_name.
- errorType: string
_validation _name requiredName of the secondary performance metric, such as mae_cv_mean or roc_auc_cv_mean.
- errorType: number
_validation _value requiredSecondary cross-validation or test-set performance value. Interpret it using error_validation_name.
- featuriserType: stringrequired
Human-readable name of the selected molecular representation, including the foundation model name when one was selected.
- isType: boolean
_log requiredWhether the model was trained on log-transformed targets.
- testType: numbernullable
_r2 R-squared on user-provided external test data, if any.