You are viewing documentation about an older version (1.0.9). View latest version

snowflake.ml.modeling.linear_model.TweedieRegressor

class snowflake.ml.modeling.linear_model.TweedieRegressor(*, power=0.0, alpha=1.0, fit_intercept=True, link='auto', solver='lbfgs', max_iter=100, tol=0.0001, warm_start=False, verbose=0, input_cols: Optional[Union[str, Iterable[str]]] = None, output_cols: Optional[Union[str, Iterable[str]]] = None, label_cols: Optional[Union[str, Iterable[str]]] = None, drop_input_cols: Optional[bool] = False, sample_weight_col: Optional[str] = None)

Bases: BaseTransformer

Generalized Linear Model with a Tweedie distribution For more details on this class, see sklearn.linear_model.TweedieRegressor

power: float, default=0

The power determines the underlying target distribution according to the following table:

Power

Distribution

0

Normal

1

Poisson

(1,2)

Compound Poisson Gamma

2

Gamma

3

Inverse Gaussian

For 0 < power < 1, no distribution exists.

alpha: float, default=1

Constant that multiplies the L2 penalty term and determines the regularization strength. alpha = 0 is equivalent to unpenalized GLMs. In this case, the design matrix X must have full column rank (no collinearities). Values of alpha must be in the range [0.0, inf).

fit_intercept: bool, default=True

Specifies if a constant (a.k.a. bias or intercept) should be added to the linear predictor (X @ coef + intercept).

link: {‘auto’, ‘identity’, ‘log’}, default=’auto’

The link function of the GLM, i.e. mapping from linear predictor X @ coeff + intercept to prediction y_pred. Option ‘auto’ sets the link depending on the chosen power parameter as follows:

  • ‘identity’ for power <= 0, e.g. for the Normal distribution

  • ‘log’ for power > 0, e.g. for Poisson, Gamma and Inverse Gaussian distributions

solver: {‘lbfgs’, ‘newton-cholesky’}, default=’lbfgs’

Algorithm to use in the optimization problem:

‘lbfgs’

Calls scipy’s L-BFGS-B optimizer.

‘newton-cholesky’

Uses Newton-Raphson steps (in arbitrary precision arithmetic equivalent to iterated reweighted least squares) with an inner Cholesky based solver. This solver is a good choice for n_samples >> n_features, especially with one-hot encoded categorical features with rare categories. Be aware that the memory usage of this solver has a quadratic dependency on n_features because it explicitly computes the Hessian matrix.

max_iter: int, default=100

The maximal number of iterations for the solver. Values must be in the range [1, inf).

tol: float, default=1e-4

Stopping criterion. For the lbfgs solver, the iteration will stop when max{|g_j|, j = 1, ..., d} <= tol where g_j is the j-th component of the gradient (derivative) of the objective function. Values must be in the range (0.0, inf).

warm_start: bool, default=False

If set to True, reuse the solution of the previous call to fit as initialization for coef_ and intercept_ .

verbose: int, default=0

For the lbfgs solver set verbose to any positive number for verbosity. Values must be in the range [0, inf).

input_cols: Optional[Union[str, List[str]]]

A string or list of strings representing column names that contain features. If this parameter is not specified, all columns in the input DataFrame except the columns specified by label_cols and sample-weight_col parameters are considered input columns.

label_cols: Optional[Union[str, List[str]]]

A string or list of strings representing column names that contain labels. This is a required param for estimators, as there is no way to infer these columns. If this parameter is not specified, then object is fitted without labels(Like a transformer).

output_cols: Optional[Union[str, List[str]]]

A string or list of strings representing column names that will store the output of predict and transform operations. The length of output_cols mus match the expected number of output columns from the specific estimator or transformer class used. If this parameter is not specified, output column names are derived by adding an OUTPUT_ prefix to the label column names. These inferred output column names work for estimator’s predict() method, but output_cols must be set explicitly for transformers.

sample_weight_col: Optional[str]

A string representing the column name containing the examples’ weights. This argument is only required when working with weighted datasets.

drop_input_cols: Optional[bool], default=False

If set, the response of predict(), transform() methods will not contain input columns.

Methods

fit(dataset)

Fit a Generalized Linear Model For more details on this function, see sklearn.linear_model.TweedieRegressor.fit

predict(dataset)

Predict using GLM with feature matrix X For more details on this function, see sklearn.linear_model.TweedieRegressor.predict

score(dataset)

Compute D^2, the percentage of deviance explained For more details on this function, see sklearn.linear_model.TweedieRegressor.score

set_input_cols(input_cols)

Input columns setter.

to_sklearn()

Get sklearn.linear_model.TweedieRegressor object.

Attributes

model_signatures

Returns model signature of current class.