Quick¶

endgame.quick.classify(X, y, preset='default', metric='roc_auc', cv_folds=None, random_state=42, verbose=True, logger=None)[source]¶

Quick classification with automatic model selection.

Parameters:

X (array-like) – Training features.
y (array-like) – Target labels.
preset (str, default='default') – Preset configuration: ‘fast’, ‘default’, ‘competition’, ‘interpretable’.
metric (str, default='roc_auc') – Scoring metric: ‘roc_auc’, ‘accuracy’, ‘f1’.
cv_folds (int, optional) – Number of CV folds. If None, uses preset default.
random_state (int, default=42) – Random seed.
verbose (bool, default=True) – Whether to print progress.
logger (ExperimentLogger, optional) – Experiment logger for tracking params and metrics.

Return type:

QuickResult

Returns:

QuickResult – Result containing model, OOF predictions, and CV score.

Examples

>>> import endgame as eg
>>> result = eg.quick.classify(X, y)
>>> print(f"CV Score: {result.cv_score:.4f}")
>>> predictions = result.model.predict(X_test)

endgame.quick.regress(X, y, preset='default', metric='rmse', cv_folds=None, random_state=42, verbose=True, logger=None)[source]¶

Quick regression with automatic model selection.

Parameters:

X (array-like) – Training features.
y (array-like) – Target values.
preset (str, default='default') – Preset configuration: ‘fast’, ‘default’, ‘competition’, ‘interpretable’.
metric (str, default='rmse') – Scoring metric: ‘rmse’, ‘r2’, ‘mae’.
cv_folds (int, optional) – Number of CV folds. If None, uses preset default.
random_state (int, default=42) – Random seed.
verbose (bool, default=True) – Whether to print progress.
logger (ExperimentLogger, optional) – Experiment logger for tracking params and metrics.

Return type:

QuickResult

Returns:

QuickResult – Result containing model, OOF predictions, and CV score.

Examples

>>> import endgame as eg
>>> result = eg.quick.regress(X, y)
>>> print(f"CV RMSE: {result.cv_score:.4f}")
>>> predictions = result.model.predict(X_test)

endgame.quick.compare(X, y, task='classification', preset='default', metric=None, cv_folds=None, random_state=42, verbose=True, logger=None)[source]¶

Compare multiple models quickly.

Parameters:

X (array-like) – Training features.
y (array-like) – Target values/labels.
task (str, default='classification') – Task type: ‘classification’ or ‘regression’.
preset (str, default='default') – Preset configuration.
metric (str, optional) – Scoring metric. If None, uses default for task.
cv_folds (int, optional) – Number of CV folds.
random_state (int, default=42) – Random seed.
verbose (bool, default=True) – Whether to print progress.
logger (ExperimentLogger, optional) – Experiment logger for tracking params and metrics.

Return type:

ComparisonResult

Returns:

ComparisonResult – Comparison results with leaderboard.

Examples

>>> import endgame as eg
>>> comparison = eg.quick.compare(X, y, task='classification')
>>> print(comparison)  # Shows leaderboard
>>> best_model = comparison.best_model

class endgame.quick.QuickResult(model, oof_predictions, cv_score, metric, feature_importances=<factory>)[source]¶

Bases: object

Result from quick.classify() or quick.regress().

Parameters:

model (Any)
oof_predictions (ndarray)
cv_score (float)
metric (str)
feature_importances (dict[str, float])

model¶

The fitted model.

Type:: Any

oof_predictions¶

Out-of-fold predictions.

Type:: ndarray

cv_score¶

Cross-validation score.

Type:: float

metric¶

Metric used for scoring.

Type:: str

feature_importances¶

Feature importance dictionary (if available).

Type:: Dict[str, float]

model: Any¶

oof_predictions: ndarray¶

cv_score: float¶

metric: str¶

feature_importances: dict[str, float]¶

class endgame.quick.ComparisonResult(results, best_model, leaderboard, metric)[source]¶

Bases: object

Result from quick.compare().

Parameters:

results (list[ModelResult])
best_model (Any)
leaderboard (list[dict[str, Any]])
metric (str)

results¶

Results for each model, sorted by score.

Type:: List[ModelResult]

best_model¶

The best performing model.

Type:: Any

leaderboard¶

Leaderboard with model names and scores.

Type:: List[Dict]

metric¶

Metric used for scoring.

Type:: str

results: list[ModelResult]¶

best_model: Any¶

leaderboard: list[dict[str, Any]]¶

metric: str¶