Source code for h2o.estimators.glm

#!/usr/bin/env python
# -*- encoding: utf-8 -*-
#
# This file is auto-generated by h2o-3/h2o-bindings/bin/gen_python.py
# Copyright 2016 H2O.ai;  Apache License Version 2.0 (see LICENSE for details)
#
from __future__ import absolute_import, division, print_function, unicode_literals

from h2o.estimators.estimator_base import H2OEstimator
from h2o.exceptions import H2OValueError
from h2o.frame import H2OFrame
from h2o.utils.typechecks import assert_is_type, Enum, numeric
import h2o


[docs]class H2OGeneralizedLinearEstimator(H2OEstimator): """ Generalized Linear Modeling Fits a generalized linear model, specified by a response variable, a set of predictors, and a description of the error distribution. A subclass of :class:`ModelBase` is returned. The specific subclass depends on the machine learning task at hand (if it's binomial classification, then an H2OBinomialModel is returned, if it's regression then a H2ORegressionModel is returned). The default print-out of the models is shown, but further GLM-specific information can be queried out of the object. Upon completion of the GLM, the resulting object has coefficients, normalized coefficients, residual/null deviance, aic, and a host of model metrics including MSE, AUC (for logistic regression), degrees of freedom, and confusion matrices. """ algo = "glm" def __init__(self, **kwargs): super(H2OGeneralizedLinearEstimator, self).__init__() self._parms = {} names_list = {"model_id", "training_frame", "validation_frame", "nfolds", "seed", "keep_cross_validation_predictions", "keep_cross_validation_fold_assignment", "fold_assignment", "fold_column", "response_column", "ignored_columns", "ignore_const_cols", "score_each_iteration", "offset_column", "weights_column", "family", "tweedie_variance_power", "tweedie_link_power", "solver", "alpha", "lambda_", "lambda_search", "early_stopping", "nlambdas", "standardize", "missing_values_handling", "compute_p_values", "remove_collinear_columns", "intercept", "non_negative", "max_iterations", "objective_epsilon", "beta_epsilon", "gradient_epsilon", "link", "prior", "lambda_min_ratio", "beta_constraints", "max_active_predictors", "interactions", "balance_classes", "class_sampling_factors", "max_after_balance_size", "max_confusion_matrix_size", "max_hit_ratio_k", "max_runtime_secs"} if "Lambda" in kwargs: kwargs["lambda_"] = kwargs.pop("Lambda") for pname, pvalue in kwargs.items(): if pname == 'model_id': self._id = pvalue self._parms["model_id"] = pvalue elif pname in names_list: # Using setattr(...) will invoke type-checking of the arguments setattr(self, pname, pvalue) else: raise H2OValueError("Unknown parameter %s = %r" % (pname, pvalue)) @property def training_frame(self): """ Id of the training data frame (Not required, to allow initial validation of model parameters). Type: ``H2OFrame``. """ return self._parms.get("training_frame") @training_frame.setter def training_frame(self, training_frame): assert_is_type(training_frame, None, H2OFrame) self._parms["training_frame"] = training_frame @property def validation_frame(self): """ Id of the validation data frame. Type: ``H2OFrame``. """ return self._parms.get("validation_frame") @validation_frame.setter def validation_frame(self, validation_frame): assert_is_type(validation_frame, None, H2OFrame) self._parms["validation_frame"] = validation_frame @property def nfolds(self): """ Number of folds for N-fold cross-validation (0 to disable or >= 2). Type: ``int`` (default: ``0``). """ return self._parms.get("nfolds") @nfolds.setter def nfolds(self, nfolds): assert_is_type(nfolds, None, int) self._parms["nfolds"] = nfolds @property def seed(self): """ Seed for pseudo random number generator (if applicable) Type: ``int`` (default: ``-1``). """ return self._parms.get("seed") @seed.setter def seed(self, seed): assert_is_type(seed, None, int) self._parms["seed"] = seed @property def keep_cross_validation_predictions(self): """ Whether to keep the predictions of the cross-validation models. Type: ``bool`` (default: ``False``). """ return self._parms.get("keep_cross_validation_predictions") @keep_cross_validation_predictions.setter def keep_cross_validation_predictions(self, keep_cross_validation_predictions): assert_is_type(keep_cross_validation_predictions, None, bool) self._parms["keep_cross_validation_predictions"] = keep_cross_validation_predictions @property def keep_cross_validation_fold_assignment(self): """ Whether to keep the cross-validation fold assignment. Type: ``bool`` (default: ``False``). """ return self._parms.get("keep_cross_validation_fold_assignment") @keep_cross_validation_fold_assignment.setter def keep_cross_validation_fold_assignment(self, keep_cross_validation_fold_assignment): assert_is_type(keep_cross_validation_fold_assignment, None, bool) self._parms["keep_cross_validation_fold_assignment"] = keep_cross_validation_fold_assignment @property def fold_assignment(self): """ Cross-validation fold assignment scheme, if fold_column is not specified. The 'Stratified' option will stratify the folds based on the response variable, for classification problems. One of: ``"auto"``, ``"random"``, ``"modulo"``, ``"stratified"`` (default: ``"auto"``). """ return self._parms.get("fold_assignment") @fold_assignment.setter def fold_assignment(self, fold_assignment): assert_is_type(fold_assignment, None, Enum("auto", "random", "modulo", "stratified")) self._parms["fold_assignment"] = fold_assignment @property def fold_column(self): """ Column with cross-validation fold index assignment per observation. Type: ``str``. """ return self._parms.get("fold_column") @fold_column.setter def fold_column(self, fold_column): assert_is_type(fold_column, None, str) self._parms["fold_column"] = fold_column @property def response_column(self): """ Response variable column. Type: ``str``. """ return self._parms.get("response_column") @response_column.setter def response_column(self, response_column): assert_is_type(response_column, None, str) self._parms["response_column"] = response_column @property def ignored_columns(self): """ Names of columns to ignore for training. Type: ``List[str]``. """ return self._parms.get("ignored_columns") @ignored_columns.setter def ignored_columns(self, ignored_columns): assert_is_type(ignored_columns, None, [str]) self._parms["ignored_columns"] = ignored_columns @property def ignore_const_cols(self): """ Ignore constant columns. Type: ``bool`` (default: ``True``). """ return self._parms.get("ignore_const_cols") @ignore_const_cols.setter def ignore_const_cols(self, ignore_const_cols): assert_is_type(ignore_const_cols, None, bool) self._parms["ignore_const_cols"] = ignore_const_cols @property def score_each_iteration(self): """ Whether to score during each iteration of model training. Type: ``bool`` (default: ``False``). """ return self._parms.get("score_each_iteration") @score_each_iteration.setter def score_each_iteration(self, score_each_iteration): assert_is_type(score_each_iteration, None, bool) self._parms["score_each_iteration"] = score_each_iteration @property def offset_column(self): """ Offset column. This will be added to the combination of columns before applying the link function. Type: ``str``. """ return self._parms.get("offset_column") @offset_column.setter def offset_column(self, offset_column): assert_is_type(offset_column, None, str) self._parms["offset_column"] = offset_column @property def weights_column(self): """ Column with observation weights. Giving some observation a weight of zero is equivalent to excluding it from the dataset; giving an observation a relative weight of 2 is equivalent to repeating that row twice. Negative weights are not allowed. Type: ``str``. """ return self._parms.get("weights_column") @weights_column.setter def weights_column(self, weights_column): assert_is_type(weights_column, None, str) self._parms["weights_column"] = weights_column @property def family(self): """ Family. Use binomial for classification with logistic regression, others are for regression problems. One of: ``"gaussian"``, ``"binomial"``, ``"quasibinomial"``, ``"multinomial"``, ``"poisson"``, ``"gamma"``, ``"tweedie"`` (default: ``"gaussian"``). """ return self._parms.get("family") @family.setter def family(self, family): assert_is_type(family, None, Enum("gaussian", "binomial", "quasibinomial", "multinomial", "poisson", "gamma", "tweedie")) self._parms["family"] = family @property def tweedie_variance_power(self): """ Tweedie variance power Type: ``float`` (default: ``0``). """ return self._parms.get("tweedie_variance_power") @tweedie_variance_power.setter def tweedie_variance_power(self, tweedie_variance_power): assert_is_type(tweedie_variance_power, None, numeric) self._parms["tweedie_variance_power"] = tweedie_variance_power @property def tweedie_link_power(self): """ Tweedie link power Type: ``float`` (default: ``1``). """ return self._parms.get("tweedie_link_power") @tweedie_link_power.setter def tweedie_link_power(self, tweedie_link_power): assert_is_type(tweedie_link_power, None, numeric) self._parms["tweedie_link_power"] = tweedie_link_power @property def solver(self): """ AUTO will set the solver based on given data and the other parameters. IRLSM is fast on on problems with small number of predictors and for lambda-search with L1 penalty, L_BFGS scales better for datasets with many columns. Coordinate descent is experimental (beta). One of: ``"auto"``, ``"irlsm"``, ``"l_bfgs"``, ``"coordinate_descent_naive"``, ``"coordinate_descent"`` (default: ``"auto"``). """ return self._parms.get("solver") @solver.setter def solver(self, solver): assert_is_type(solver, None, Enum("auto", "irlsm", "l_bfgs", "coordinate_descent_naive", "coordinate_descent")) self._parms["solver"] = solver @property def alpha(self): """ Distribution of regularization between the L1 (Lasso) and L2 (Ridge) penalties. A value of 1 for alpha represents Lasso regression, a value of 0 produces Ridge regression, and anything in between specifies the amount of mixing between the two. Default value of alpha is 0 when SOLVER = 'L-BFGS'; 0.5 otherwise. Type: ``List[float]``. """ return self._parms.get("alpha") @alpha.setter def alpha(self, alpha): assert_is_type(alpha, None, numeric, [numeric]) self._parms["alpha"] = alpha @property def lambda_(self): """ Regularization strength Type: ``List[float]``. """ return self._parms.get("lambda") @lambda_.setter def lambda_(self, lambda_): assert_is_type(lambda_, None, numeric, [numeric]) self._parms["lambda"] = lambda_ @property def lambda_search(self): """ Use lambda search starting at lambda max, given lambda is then interpreted as lambda min Type: ``bool`` (default: ``False``). """ return self._parms.get("lambda_search") @lambda_search.setter def lambda_search(self, lambda_search): assert_is_type(lambda_search, None, bool) self._parms["lambda_search"] = lambda_search @property def early_stopping(self): """ Stop early when there is no more relative improvement on train or validation (if provided) Type: ``bool`` (default: ``True``). """ return self._parms.get("early_stopping") @early_stopping.setter def early_stopping(self, early_stopping): assert_is_type(early_stopping, None, bool) self._parms["early_stopping"] = early_stopping @property def nlambdas(self): """ Number of lambdas to be used in a search. Default indicates: If alpha is zero, with lambda search set to True, the value of nlamdas is set to 30 (fewer lambdas are needed for ridge regression) otherwise it is set to 100. Type: ``int`` (default: ``-1``). """ return self._parms.get("nlambdas") @nlambdas.setter def nlambdas(self, nlambdas): assert_is_type(nlambdas, None, int) self._parms["nlambdas"] = nlambdas @property def standardize(self): """ Standardize numeric columns to have zero mean and unit variance Type: ``bool`` (default: ``True``). """ return self._parms.get("standardize") @standardize.setter def standardize(self, standardize): assert_is_type(standardize, None, bool) self._parms["standardize"] = standardize @property def missing_values_handling(self): """ Handling of missing values. Either MeanImputation or Skip. One of: ``"mean_imputation"``, ``"skip"`` (default: ``"mean_imputation"``). """ return self._parms.get("missing_values_handling") @missing_values_handling.setter def missing_values_handling(self, missing_values_handling): assert_is_type(missing_values_handling, None, Enum("mean_imputation", "skip")) self._parms["missing_values_handling"] = missing_values_handling @property def compute_p_values(self): """ Request p-values computation, p-values work only with IRLSM solver and no regularization Type: ``bool`` (default: ``False``). """ return self._parms.get("compute_p_values") @compute_p_values.setter def compute_p_values(self, compute_p_values): assert_is_type(compute_p_values, None, bool) self._parms["compute_p_values"] = compute_p_values @property def remove_collinear_columns(self): """ In case of linearly dependent columns, remove some of the dependent columns Type: ``bool`` (default: ``False``). """ return self._parms.get("remove_collinear_columns") @remove_collinear_columns.setter def remove_collinear_columns(self, remove_collinear_columns): assert_is_type(remove_collinear_columns, None, bool) self._parms["remove_collinear_columns"] = remove_collinear_columns @property def intercept(self): """ Include constant term in the model Type: ``bool`` (default: ``True``). """ return self._parms.get("intercept") @intercept.setter def intercept(self, intercept): assert_is_type(intercept, None, bool) self._parms["intercept"] = intercept @property def non_negative(self): """ Restrict coefficients (not intercept) to be non-negative Type: ``bool`` (default: ``False``). """ return self._parms.get("non_negative") @non_negative.setter def non_negative(self, non_negative): assert_is_type(non_negative, None, bool) self._parms["non_negative"] = non_negative @property def max_iterations(self): """ Maximum number of iterations Type: ``int`` (default: ``-1``). """ return self._parms.get("max_iterations") @max_iterations.setter def max_iterations(self, max_iterations): assert_is_type(max_iterations, None, int) self._parms["max_iterations"] = max_iterations @property def objective_epsilon(self): """ Converge if objective value changes less than this. Default indicates: If lambda_search is set to True the value of objective_epsilon is set to .0001. If the lambda_search is set to False and lambda is equal to zero, the value of objective_epsilon is set to .000001, for any other value of lambda the default value of objective_epsilon is set to .0001. Type: ``float`` (default: ``-1``). """ return self._parms.get("objective_epsilon") @objective_epsilon.setter def objective_epsilon(self, objective_epsilon): assert_is_type(objective_epsilon, None, numeric) self._parms["objective_epsilon"] = objective_epsilon @property def beta_epsilon(self): """ Converge if beta changes less (using L-infinity norm) than beta esilon, ONLY applies to IRLSM solver Type: ``float`` (default: ``0.0001``). """ return self._parms.get("beta_epsilon") @beta_epsilon.setter def beta_epsilon(self, beta_epsilon): assert_is_type(beta_epsilon, None, numeric) self._parms["beta_epsilon"] = beta_epsilon @property def gradient_epsilon(self): """ Converge if objective changes less (using L-infinity norm) than this, ONLY applies to L-BFGS solver. Default indicates: If lambda_search is set to False and lambda is equal to zero, the default value of gradient_epsilon is equal to .000001, otherwise the default value is .0001. If lambda_search is set to True, the conditional values above are 1E-8 and 1E-6 respectively. Type: ``float`` (default: ``-1``). """ return self._parms.get("gradient_epsilon") @gradient_epsilon.setter def gradient_epsilon(self, gradient_epsilon): assert_is_type(gradient_epsilon, None, numeric) self._parms["gradient_epsilon"] = gradient_epsilon @property def link(self): """ One of: ``"family_default"``, ``"identity"``, ``"logit"``, ``"log"``, ``"inverse"``, ``"tweedie"`` (default: ``"family_default"``). """ return self._parms.get("link") @link.setter def link(self, link): assert_is_type(link, None, Enum("family_default", "identity", "logit", "log", "inverse", "tweedie")) self._parms["link"] = link @property def prior(self): """ Prior probability for y==1. To be used only for logistic regression iff the data has been sampled and the mean of response does not reflect reality. Type: ``float`` (default: ``-1``). """ return self._parms.get("prior") @prior.setter def prior(self, prior): assert_is_type(prior, None, numeric) self._parms["prior"] = prior @property def lambda_min_ratio(self): """ Minimum lambda used in lambda search, specified as a ratio of lambda_max. Default indicates: if the number of observations is greater than the number of variables then lambda_min_ratio is set to 0.0001; if the number of observations is less than the number of variables then lambda_min_ratio is set to 0.01. Type: ``float`` (default: ``-1``). """ return self._parms.get("lambda_min_ratio") @lambda_min_ratio.setter def lambda_min_ratio(self, lambda_min_ratio): assert_is_type(lambda_min_ratio, None, numeric) self._parms["lambda_min_ratio"] = lambda_min_ratio @property def beta_constraints(self): """ Beta constraints Type: ``H2OFrame``. """ return self._parms.get("beta_constraints") @beta_constraints.setter def beta_constraints(self, beta_constraints): assert_is_type(beta_constraints, None, H2OFrame) self._parms["beta_constraints"] = beta_constraints @property def max_active_predictors(self): """ Maximum number of active predictors during computation. Use as a stopping criterion to prevent expensive model building with many predictors. Default indicates: If the IRLSM solver is used, the value of max_active_predictors is set to 7000 otherwise it is set to 100000000. Type: ``int`` (default: ``-1``). """ return self._parms.get("max_active_predictors") @max_active_predictors.setter def max_active_predictors(self, max_active_predictors): assert_is_type(max_active_predictors, None, int) self._parms["max_active_predictors"] = max_active_predictors @property def interactions(self): """ A list of predictor column indices to interact. All pairwise combinations will be computed for the list. Type: ``List[str]``. """ return self._parms.get("interactions") @interactions.setter def interactions(self, interactions): assert_is_type(interactions, None, [str]) self._parms["interactions"] = interactions @property def balance_classes(self): """ Balance training data class counts via over/under-sampling (for imbalanced data). Type: ``bool`` (default: ``False``). """ return self._parms.get("balance_classes") @balance_classes.setter def balance_classes(self, balance_classes): assert_is_type(balance_classes, None, bool) self._parms["balance_classes"] = balance_classes @property def class_sampling_factors(self): """ Desired over/under-sampling ratios per class (in lexicographic order). If not specified, sampling factors will be automatically computed to obtain class balance during training. Requires balance_classes. Type: ``List[float]``. """ return self._parms.get("class_sampling_factors") @class_sampling_factors.setter def class_sampling_factors(self, class_sampling_factors): assert_is_type(class_sampling_factors, None, [float]) self._parms["class_sampling_factors"] = class_sampling_factors @property def max_after_balance_size(self): """ Maximum relative size of the training data after balancing class counts (can be less than 1.0). Requires balance_classes. Type: ``float`` (default: ``5``). """ return self._parms.get("max_after_balance_size") @max_after_balance_size.setter def max_after_balance_size(self, max_after_balance_size): assert_is_type(max_after_balance_size, None, float) self._parms["max_after_balance_size"] = max_after_balance_size @property def max_confusion_matrix_size(self): """ [Deprecated] Maximum size (# classes) for confusion matrices to be printed in the Logs Type: ``int`` (default: ``20``). """ return self._parms.get("max_confusion_matrix_size") @max_confusion_matrix_size.setter def max_confusion_matrix_size(self, max_confusion_matrix_size): assert_is_type(max_confusion_matrix_size, None, int) self._parms["max_confusion_matrix_size"] = max_confusion_matrix_size @property def max_hit_ratio_k(self): """ Maximum number (top K) of predictions to use for hit ratio computation (for multi-class only, 0 to disable) Type: ``int`` (default: ``0``). """ return self._parms.get("max_hit_ratio_k") @max_hit_ratio_k.setter def max_hit_ratio_k(self, max_hit_ratio_k): assert_is_type(max_hit_ratio_k, None, int) self._parms["max_hit_ratio_k"] = max_hit_ratio_k @property def max_runtime_secs(self): """ Maximum allowed runtime in seconds for model training. Use 0 to disable. Type: ``float`` (default: ``0``). """ return self._parms.get("max_runtime_secs") @max_runtime_secs.setter def max_runtime_secs(self, max_runtime_secs): assert_is_type(max_runtime_secs, None, numeric) self._parms["max_runtime_secs"] = max_runtime_secs @property def Lambda(self): """DEPRECATED. Use ``self.lambda_`` instead""" return self._parms["lambda"] if "lambda" in self._parms else None @Lambda.setter def Lambda(self, value): self._parms["lambda"] = value @staticmethod
[docs] def getGLMRegularizationPath(model): """ Extract full regularization path explored during lambda search from glm model. :param model: source lambda search model """ x = h2o.api("GET /3/GetGLMRegPath", data={"model": model._model_json["model_id"]["name"]}) ns = x.pop("coefficient_names") res = { "lambdas": x["lambdas"], "explained_deviance_train": x["explained_deviance_train"], "explained_deviance_valid": x["explained_deviance_valid"], "coefficients": [dict(zip(ns, y)) for y in x["coefficients"]], } if "coefficients_std" in x: res["coefficients_std"] = [dict(zip(ns, y)) for y in x["coefficients_std"]] return res
@staticmethod
[docs] def makeGLMModel(model, coefs, threshold=.5): """ Create a custom GLM model using the given coefficients. Needs to be passed source model trained on the dataset to extract the dataset information from. :param model: source model, used for extracting dataset information :param coefs: dictionary containing model coefficients :param threshold: (optional, only for binomial) decision threshold used for classification """ model_json = h2o.api( "POST /3/MakeGLMModel", data={"model": model._model_json["model_id"]["name"], "names": list(coefs.keys()), "beta": list(coefs.values()), "threshold": threshold} ) m = H2OGeneralizedLinearEstimator() m._resolve_model(model_json["model_id"]["name"], model_json) return m