Module panama.ml.tunable.regression

Classes

class TunableCatBoostRegressor (name: str = 'catboost_regr', random_state: int = 42)

A class for a tunable CatBoost regressor model.

Initializes the TunableCatBoostRegressor with a name and a CatBoostRegressor model.

Args

name

The name of the model.

Returns

None.

Expand source code

class TunableCatBoostRegressor(BaseTunableModel):
    """A class for a tunable CatBoost regressor model."""

    DEFAULT_SEARCH_DICT = {
        "iterations": {"type": "int", "min": 100, "max": 1000},
        "depth": {"type": "int", "min": 4, "max": 12},
        "l2_leaf_reg": {"type": "choice", "values": [0.1, 0.2, 0.5, 1, 3, 4, 5]},
        "learning_rate": {"type": "float", "min": 0.001, "max": 0.5},
        "border_count": {"type": "int", "min": 1, "max": 255},
        "loss_function": {"type": "choice", "values": ["RMSE", "MAE", "MAPE"]},
    }

    def __init__(self, name: str = "catboost_regr", random_state: int = 42):
        """Initializes the TunableCatBoostRegressor with a name and a CatBoostRegressor model.

        Args:
            name: The name of the model.

        Returns:
            None.
        """
        self.name = name
        self.model = CatBoostRegressor(random_state=random_state)

    def fit(self, X: DataFrame, y: Union[DataFrame, Series]) -> None:
        """Fits the model to the training data.

        Args:
            X: The features for training.
            y: The target values for training.

        Returns:
            None.
        """
        self.model.fit(X, y)

    def predict(self, X: DataFrame) -> List:
        """Predicts target values for the given features.

        Args:
            X: The features to predict target values for.

        Returns:
            List: The predicted target values.
        """
        return self.model.predict(X)

    def set_params(self, params: Dict) -> None:
        """Sets the hyperparameters of the model.

        Args:
            params: A dictionary of hyperparameters.

        Returns:
            None.
        """
        self.model.set_params(**params)

    def get_params(self, deep: bool = True) -> Dict:
        """Set the hyperparameters of the model.

        Args:
            params: hyperparameters and their values.

        Returns:
            None.
        """
        return self.model.get_params(deep)

Ancestors

• BaseTunableModel
• abc.ABC

Class variables

var DEFAULT_SEARCH_DICT

Methods

def fit(self, X: pandas.core.frame.DataFrame, y: Union[pandas.core.frame.DataFrame, pandas.core.series.Series]) ‑> None

Fits the model to the training data.

Args

X

The features for training.

y

The target values for training.

Returns

None.

def predict(self, X: pandas.core.frame.DataFrame) ‑> List

Predicts target values for the given features.

Args

X

The features to predict target values for.

Returns

List

The predicted target values.

def set_params(self, params: Dict) ‑> None

Sets the hyperparameters of the model.

Args

params

A dictionary of hyperparameters.

Returns

None.

Inherited members

• BaseTunableModel:
  • get_model
  • get_name
  • get_params

class TunableLGBMRegressor (name: str = 'lgbm_regr', random_state: int = 42)

A base class for tunable machine learning models.

This class defines a set of abstract methods that must be implemented by any concrete subclass.

Attributes

None. Initializes the TunableLGBMRegressor with a name and a LightGBMRegressor model.

Args

name

The name of the model.

Returns

None.

Expand source code

class TunableLGBMRegressor(BaseTunableModel):
    DEFAULT_SEARCH_DICT = {
        "n_estimators": {"type": "int", "min": 100, "max": 1000},
        "max_depth": {"type": "int", "min": 4, "max": 12},
        "learning_rate": {"type": "float", "min": 0.001, "max": 0.5},
        "subsample": {"type": "float", "min": 0.3, "max": 1},
        "min_child_samples": {"type": "int", "min": 10, "max": 75},
        "num_leaves": {"type": "int", "min": 10, "max": 75},
        "boosting_type": {"type": "choice", "values": ["gbdt", "dart", "goss"]},
        "objective": {"type": "choice", "values": ["regression", "regression_l1", "mape"]},
    }
    """A class for a tunable LightGBM regressor model."""

    def __init__(self, name: str = "lgbm_regr", random_state: int = 42):
        """Initializes the TunableLGBMRegressor with a name and a LightGBMRegressor model.

        Args:
            name: The name of the model.

        Returns:
            None.
        """
        self.name = name
        self.model = LGBMRegressor(random_state=random_state)

    def fit(self, X: DataFrame, y: Union[DataFrame, Series]) -> None:
        """Fits the model to the training data.

        Args:
            X: The features for training.
            y: The target values for training.

        Returns:
            None.
        """
        self.model.fit(X, y)

    def predict(self, X: DataFrame) -> List:
        """Predicts target values for the given features.

        Args:
            X: The features to predict target values for.

        Returns:
            List: The predicted target values.
        """
        return self.model.predict(X)

    def set_params(self, params: Dict) -> None:
        """Sets the hyperparameters of the model.

        Args:
            params: A dictionary of hyperparameters.

        Returns:
            None.
        """
        self.model.set_params(**params)

    def get_params(self, deep: bool = True) -> Dict:
        """Set the hyperparameters of the model.

        Args:
            params: hyperparameters and their values.

        Returns:
            None.
        """
        return self.model.get_params(deep)

Ancestors

• BaseTunableModel
• abc.ABC

Class variables

var DEFAULT_SEARCH_DICT

A class for a tunable LightGBM regressor model.

Methods

def fit(self, X: pandas.core.frame.DataFrame, y: Union[pandas.core.frame.DataFrame, pandas.core.series.Series]) ‑> None

Fits the model to the training data.

Args

X

The features for training.

y

The target values for training.

Returns

None.

def predict(self, X: pandas.core.frame.DataFrame) ‑> List

Predicts target values for the given features.

Args

X

The features to predict target values for.

Returns

List

The predicted target values.

def set_params(self, params: Dict) ‑> None

Sets the hyperparameters of the model.

Args

params

A dictionary of hyperparameters.

Returns

None.

Inherited members

• BaseTunableModel:
  • get_model
  • get_name
  • get_params

class TunableRandomForestRegressor (name: str = 'random_forest_regr', random_state: int = 42)

A class for a tunable random forest regressor model.

Initializes the TunableRandomForestRegressor with a name and a RandomForestRegressor model.

Args

name : str

The name of the model.

Returns

None.

Expand source code

class TunableRandomForestRegressor(BaseTunableModel):
    """A class for a tunable random forest regressor model."""

    DEFAULT_SEARCH_DICT = {
        "n_estimators": {"type": "int", "min": 100, "max": 1000},
        "max_depth": {"type": "int", "min": 4, "max": 12},
        "max_features": {"type": "float", "min": 0.3, "max": 1},
        "min_samples_leaf": {"type": "int", "min": 3, "max": 25},
        "criterion": {"type": "choice", "values": ["poisson", "absolute_error", "friedman_mse", "squared_error"]},
        "bootstrap": {"type": "choice", "values": [True, False]},
    }

    def __init__(self, name: str = "random_forest_regr", random_state: int = 42):
        """Initializes the TunableRandomForestRegressor with a name and a RandomForestRegressor model.

        Args:
            name (str): The name of the model.

        Returns:
            None.
        """
        self.name = name
        self.model = RandomForestRegressor(random_state=random_state)

    def fit(self, X: DataFrame, y: Union[DataFrame, Series]) -> None:
        """Fits the model to the training data.

        Args:
            X: The features for training.
            y: The target values for training.

        Returns:
            None.
        """
        self.model.fit(X, y)

    def predict(self, X: DataFrame) -> List:
        """Predicts target values for the given features.

        Args:
            X: The features to predict target values for.

        Returns:
            List: The predicted target values.
        """
        return self.model.predict(X)

    def set_params(self, params: Dict) -> None:
        """Sets the hyperparameters of the model.

        Args:
            params: A dictionary of hyperparameters.

        Returns:
            None.
        """
        self.model.set_params(**params)

    def get_params(self, deep: bool = True) -> Dict:
        """Set the hyperparameters of the model.

        Args:
            params: hyperparameters and their values.

        Returns:
            None.
        """
        return self.model.get_params(deep)

Ancestors

• BaseTunableModel
• abc.ABC

Class variables

var DEFAULT_SEARCH_DICT

Methods

def fit(self, X: pandas.core.frame.DataFrame, y: Union[pandas.core.frame.DataFrame, pandas.core.series.Series]) ‑> None

Fits the model to the training data.

Args

X

The features for training.

y

The target values for training.

Returns

None.

def predict(self, X: pandas.core.frame.DataFrame) ‑> List

Predicts target values for the given features.

Args

X

The features to predict target values for.

Returns

List

The predicted target values.

def set_params(self, params: Dict) ‑> None

Sets the hyperparameters of the model.

Args

params

A dictionary of hyperparameters.

Returns

None.

Inherited members

• BaseTunableModel:
  • get_model
  • get_name
  • get_params

class TunableSGDRegressor (name: str = 'sgd_regr', random_state: int = 42)

A class for a tunable SGD regressor model.

Initializes the TunableSGDRegressor with a name and an SGDRegressor model.

Args

name

The name of the model.

Returns

None.

Expand source code

class TunableSGDRegressor(BaseTunableModel):
    """A class for a tunable SGD regressor model."""

    DEFAULT_SEARCH_DICT = {
        "loss": {
            "type": "choice",
            "values": ["squared_error", "huber", "epsilon_insensitive", "squared_epsilon_insensitive"],
        },
        "penalty": {"type": "choice", "values": ["l1", "l2", "elasticnet"]},
        "max_iter": {"type": "int", "min": 100, "max": 1000},
        "learning_rate": {"type": "choice", "values": ["constant", "invscaling", "optimal", "adaptive"]},
    }

    def __init__(self, name: str = "sgd_regr", random_state: int = 42):
        """Initializes the TunableSGDRegressor with a name and an SGDRegressor model.

        Args:
            name: The name of the model.

        Returns:
            None.
        """
        self.name = name
        self.model = SGDRegressor(random_state=random_state)

    def fit(self, X: DataFrame, y: Union[DataFrame, Series]) -> None:
        """Fits the model to the training data.

        Args:
            X: The features for training.
            y: The target values for training.

        Returns:
            None.
        """
        self.model.fit(X, y)

    def predict(self, X: DataFrame) -> List:
        """Predicts target values for the given features.

        Args:
            X: The features to predict target values for.

        Returns:
            List: The predicted target values.
        """
        return self.model.predict(X)

    def set_params(self, params: Dict) -> None:
        """Sets the hyperparameters of the model.

        Args:
            params: A dictionary of hyperparameters.

        Returns:
            None.
        """
        self.model.set_params(**params)

    def get_params(self, deep: bool = True) -> Dict:
        """Set the hyperparameters of the model.

        Args:
            params: hyperparameters and their values.

        Returns:
            None.
        """
        return self.model.get_params(deep)

Ancestors

• BaseTunableModel
• abc.ABC

Class variables

var DEFAULT_SEARCH_DICT

Methods

def fit(self, X: pandas.core.frame.DataFrame, y: Union[pandas.core.frame.DataFrame, pandas.core.series.Series]) ‑> None

Fits the model to the training data.

Args

X

The features for training.

y

The target values for training.

Returns

None.

def predict(self, X: pandas.core.frame.DataFrame) ‑> List

Predicts target values for the given features.

Args

X

The features to predict target values for.

Returns

List

The predicted target values.

def set_params(self, params: Dict) ‑> None

Sets the hyperparameters of the model.

Args

params

A dictionary of hyperparameters.

Returns

None.

Inherited members

• BaseTunableModel:
  • get_model
  • get_name
  • get_params

class TunableXGBRegressor (name: str = 'xgb_regr', random_state: int = 42)

A class for a tunable XGBoost model.

Initializes the TunableXgbModel with a name and an XGBRegressor model.

Args

name

The name of the model.

Returns

None.

Expand source code

class TunableXGBRegressor(BaseTunableModel):
    """A class for a tunable XGBoost model."""

    DEFAULT_SEARCH_DICT = {
        "n_estimators": {"type": "int", "min": 100, "max": 1000},
        "max_depth": {"type": "int", "min": 4, "max": 12},
        "learning_rate": {"type": "float", "min": 0.001, "max": 0.5},
        "objective": {"type": "choice", "values": ["reg:squarederror", "reg:squaredlogerror"]},
        "subsample": {"type": "float", "min": 0.3, "max": 1},
        "booster": {"type": "choice", "values": ["gbtree", "gblinear", "dart"]},
        "tree_method": {"type": "choice", "values": ["exact", "approx", "hist"]},
    }

    def __init__(self, name: str = "xgb_regr", random_state: int = 42):
        """Initializes the TunableXgbModel with a name and an XGBRegressor model.

        Args:
            name: The name of the model.

        Returns:
            None.
        """
        self.name = name
        self.model = XGBRegressor(random_state=random_state)

    def fit(self, X: DataFrame, y: Union[DataFrame, Series]) -> None:
        """Fits the model to the training data.

        Args:
            X: The features for training.
            y: The target values for training.

        Returns:
            None.
        """
        self.model.fit(X, y)

    def predict(self, X: DataFrame) -> List:
        """Predicts target values for the given features.

        Args:
            X: The features to predict target values for.

        Returns:
            List: The predicted target values.
        """
        return self.model.predict(X)

    def get_params(self, deep: bool = True) -> Dict:
        """Set the hyperparameters of the model.

        Args:
            params: hyperparameters and their values.

        Returns:
            None.
        """
        return self.model.get_params(deep)

    def set_params(self, params: Dict) -> None:
        """Sets the hyperparameters of the model.

        Args:
            params (Dict): A dictionary of hyperparameters.

        Returns:
            None.
        """
        self.model.set_params(**params)

Ancestors

• BaseTunableModel
• abc.ABC

Class variables

var DEFAULT_SEARCH_DICT

Methods

def fit(self, X: pandas.core.frame.DataFrame, y: Union[pandas.core.frame.DataFrame, pandas.core.series.Series]) ‑> None

Fits the model to the training data.

Args

X

The features for training.

y

The target values for training.

Returns

None.

def predict(self, X: pandas.core.frame.DataFrame) ‑> List

Predicts target values for the given features.

Args

X

The features to predict target values for.

Returns

List

The predicted target values.

def set_params(self, params: Dict) ‑> None

Sets the hyperparameters of the model.

Args

params : Dict

A dictionary of hyperparameters.

Returns

None.

Inherited members

• BaseTunableModel:
  • get_model
  • get_name
  • get_params