← Back to API reference

neuralbench.transforms.PredefinedSplit

pydantic model neuralbench.transforms.PredefinedSplit

Assign train/test labels based on a predefined split, and optionally split train into validation as well.

Parameters:

• event_type (str | None) – If provided, only split events of this type.

• test_split_query (str | None) – If provided, query used to create a test split.

• col_name (str) – Column name to use for the created split.

• valid_split_ratio (float) – Ratio of the training set to use for validation. CAUTION: This is unlike the other splitter (e.g. SklearnSplit) where the validation split is a ratio of the entire dataset.

• test_random_state (int | None) – Unused - for compatibility with SklearnSplit.

Fields:

• col_name (str)

• event_type (str | None)

• test_random_state (int | None)

• test_split_query (str | None)

• valid_random_state (int)

• valid_split_by (str | None)

• valid_split_ratio (float)

field event_type: str | None = None

field test_split_query: str | None [Required]

field col_name: str = 'split'

field valid_split_by: str | None = 'timeline'

field valid_split_ratio: float = 0.2

field valid_random_state: int = 33

field test_random_state: int | None = None

requirements: tp.ClassVar[tuple[str, ...]] = ()

← Back to API reference