load_credit_default
def load_credit_default(*, include_demographics: bool = False, cache_dir: str | os.PathLike | None = None, download: bool = True) -> tuple[numpy.ndarray, numpy.ndarray, list[str]]
from compileml.datasets import load_credit_default
Load the UCI credit-card default panel.
Parameters#
| Name | Type | Default | Kind |
|---|---|---|---|
include_demographicsInclude SEX, EDUCATION, MARRIAGE and AGE. Excluded by default; see the module docstring for why. | bool | False | keyword-only |
cache_dirWhere to keep the downloaded file. Defaults to $COMPILEML_DATA_HOME or ~/.compileml/data. | str | os.PathLike | None | None | keyword-only |
| downloadWhen False, raise instead of reaching the network if the cache is cold. | bool | True | keyword-only |
Returns#
tuple[numpy.ndarray, numpy.ndarray, list[str]]
(X, y, feature_names) — X float64 of shape (30000, 19), or
(30000, 23) with demographics; y int of shape (30000,) with
a 22.12% base rate; feature_names in column order.
Raises#
- FileNotFoundError
- OSError
Read from the function body and the private helpers it calls, not inferred. A function that raises nothing has no section here.
Notes#
Every value in the source is an integer, including the currency amounts
(New Taiwan dollars, unrounded). X is returned as float64 because
that is what the compile side expects, but no precision is lost in the
conversion, and the split comparisons the artifact performs are exact.