Log Grid Search

class abil.log_grid_search.LogGridSearch(m, verbose, cv, param_grid, scoring, regions=None)

Bases: object

Perform grid search with optional logarithmic transformation of the target variable.

Supports evaluating models with no transformation, log-transformation, or both.

do_exp(x)

Apply exponential transformation to the input values.

Parameters:

x (array-like) – Input values.

Returns:

y – Exponentially transformed values.

Return type:

array-like

do_log(x)

Apply natural logarithm transformation to the input values.

Parameters:

x (array-like) – Input values.

Returns:

y – Log-transformed values.

Return type:

array-like

do_nothing(x)

Apply no transformation to the input values.

Parameters:

x (array-like) – Input values.

Returns:

y – Non-transformed values.

Return type:

array-like

transformed_fit(X, y, log, predictors)

Perform grid search with optional log transformation on the target variable.

Parameters:

• X (pd.DataFrame or np.ndarray) – Feature matrix.

• y (pd.Series or np.ndarray) – Target variable.

• log (str) – Transformation mode: “yes” (log transform), “no” (no transform), or “both” (test both).

• predictors (list) – Predictors used for training (not directly used in the function).

Returns:

Fitted grid search instance for the best-performing model.

Return type:

GridSearchCV

Notes

• Applies log transformation using TransformedTargetRegressor when log=”yes”.

• If log=”both”, compares models with and without log transformation.

• Uses self.param_grid, self.scoring, and self.cv for grid search.

Raises:

ValueError – If log is not “yes”, “no”, or “both”.