# Copyright (c) PyWhy contributors. All rights reserved.
# Licensed under the MIT License.
#
# This code contains snippets of code from
# https://github.com/scikit-learn/scikit-learn/blob/main/sklearn/ensemble/_forest.py
# published under the following license and copyright:
# BSD 3-Clause License
#
# Copyright (c) 2007-2020 The scikit-learn developers.
# All rights reserved.
import numbers
from warnings import catch_warnings, simplefilter, warn
from abc import ABCMeta, abstractmethod
import numpy as np
import threading
from ..._ensemble import (BaseEnsemble, _partition_estimators, _get_n_samples_subsample, _accumulate_prediction)
from ...utilities import check_inputs, cross_product
from ...tree._tree import DTYPE, DOUBLE
from ._tree import PolicyTree
from joblib import Parallel, delayed
from scipy.sparse import hstack as sparse_hstack
from sklearn.utils import check_random_state, compute_sample_weight
from sklearn.utils.validation import _check_sample_weight, check_is_fitted
from sklearn.utils import check_X_y
__all__ = ["PolicyForest"]
MAX_INT = np.iinfo(np.int32).max
# =============================================================================
# Policy Forest
# =============================================================================
[docs]class PolicyForest(BaseEnsemble, metaclass=ABCMeta):
""" Welfare maximization policy forest. Trains a forest to maximize the objective:
:math:`1/n \\sum_i \\sum_j a_j(X_i) * y_{ij}`, where, where :math:`a(X)` is constrained
to take value of 1 only on one coordinate and zero otherwise. This corresponds to a policy
optimization problem.
Parameters
----------
n_estimators : int, default 100
The total number of trees in the forest. The forest consists of a
forest of sqrt(n_estimators) sub-forests, where each sub-forest
contains sqrt(n_estimators) trees.
criterion : {``'neg_welfare'``}, default 'neg_welfare'
The criterion type
max_depth : int, default None
The maximum depth of the tree. If None, then nodes are expanded until
all leaves are pure or until all leaves contain less than
min_samples_split samples.
min_samples_split : int or float, default 10
The minimum number of samples required to split an internal node:
- If int, then consider `min_samples_split` as the minimum number.
- If float, then `min_samples_split` is a fraction and
`ceil(min_samples_split * n_samples)` are the minimum
number of samples for each split.
min_samples_leaf : int or float, default 5
The minimum number of samples required to be at a leaf node.
A split point at any depth will only be considered if it leaves at
least ``min_samples_leaf`` training samples in each of the left and
right branches. This may have the effect of smoothing the model,
especially in regression.
- If int, then consider `min_samples_leaf` as the minimum number.
- If float, then `min_samples_leaf` is a fraction and
`ceil(min_samples_leaf * n_samples)` are the minimum
number of samples for each node.
min_weight_fraction_leaf : float, default 0.0
The minimum weighted fraction of the sum total of weights (of all
the input samples) required to be at a leaf node. Samples have
equal weight when sample_weight is not provided.
max_features : int, float, {"auto", "sqrt", "log2"}, or None, default None
The number of features to consider when looking for the best split:
- If int, then consider `max_features` features at each split.
- If float, then `max_features` is a fraction and
`int(max_features * n_features)` features are considered at each
split.
- If "auto", then `max_features=n_features`.
- If "sqrt", then `max_features=sqrt(n_features)`.
- If "log2", then `max_features=log2(n_features)`.
- If None, then `max_features=n_features`.
Note: the search for a split does not stop until at least one
valid partition of the node samples is found, even if it requires to
effectively inspect more than ``max_features`` features.
min_impurity_decrease : float, default 0.0
A node will be split if this split induces a decrease of the impurity
greater than or equal to this value.
The weighted impurity decrease equation is the following::
N_t / N * (impurity - N_t_R / N_t * right_impurity
- N_t_L / N_t * left_impurity)
where ``N`` is the total number of samples, ``N_t`` is the number of
samples at the current node, ``N_t_L`` is the number of samples in the
left child, and ``N_t_R`` is the number of samples in the right child.
``N``, ``N_t``, ``N_t_R`` and ``N_t_L`` all refer to the weighted sum,
if ``sample_weight`` is passed.
max_samples : int or float in (0, 1], default .5,
The number of samples to use for each subsample that is used to train each tree:
- If int, then train each tree on `max_samples` samples, sampled without replacement from all the samples
- If float, then train each tree on ceil(`max_samples` * `n_samples`), sampled without replacement
from all the samples.
min_balancedness_tol: float in [0, .5], default .45
How imbalanced a split we can tolerate. This enforces that each split leaves at least
(.5 - min_balancedness_tol) fraction of samples on each side of the split; or fraction
of the total weight of samples, when sample_weight is not None. Default value, ensures
that at least 5% of the parent node weight falls in each side of the split. Set it to 0.0 for no
balancedness and to .5 for perfectly balanced splits. For the formal inference theory
to be valid, this has to be any positive constant bounded away from zero.
honest: bool, default True
Whether the data should be split in two equally sized samples, such that the one half-sample
is used to determine the optimal split at each node and the other sample is used to determine
the value of every node.
n_jobs : int or None, default -1
The number of jobs to run in parallel for both `fit` and `predict`.
``None`` means 1 unless in a :func:`joblib.parallel_backend` context.
``-1`` means using all processors. See :term:`Glossary <n_jobs>`
for more details.
verbose : int, default 0
Controls the verbosity when fitting and predicting.
random_state : int, RandomState instance, or None, default None
If int, random_state is the seed used by the random number generator;
If :class:`~numpy.random.mtrand.RandomState` instance, random_state is the random number generator;
If None, the random number generator is the :class:`~numpy.random.mtrand.RandomState` instance used
by :mod:`np.random<numpy.random>`.
warm_start : bool, default False
When set to ``True``, reuse the solution of the previous call to fit
and add more estimators to the ensemble, otherwise, just fit a whole
new forest.
Attributes
----------
feature_importances_ : ndarray of shape (n_features,)
The feature importances based on the amount of parameter heterogeneity they create.
The higher, the more important the feature.
"""
[docs] def __init__(self,
n_estimators=100, *,
criterion='neg_welfare',
max_depth=None,
min_samples_split=10,
min_samples_leaf=5,
min_weight_fraction_leaf=0.,
max_features="auto",
min_impurity_decrease=0.,
max_samples=.5,
min_balancedness_tol=.45,
honest=True,
n_jobs=-1,
random_state=None,
verbose=0,
warm_start=False):
super().__init__(
base_estimator=PolicyTree(),
n_estimators=n_estimators,
estimator_params=("criterion", "max_depth", "min_samples_split",
"min_samples_leaf", "min_weight_fraction_leaf",
"max_features", "min_impurity_decrease", "honest",
"min_balancedness_tol",
"random_state"))
self.criterion = criterion
self.max_depth = max_depth
self.min_samples_split = min_samples_split
self.min_samples_leaf = min_samples_leaf
self.min_weight_fraction_leaf = min_weight_fraction_leaf
self.max_features = max_features
self.min_impurity_decrease = min_impurity_decrease
self.min_balancedness_tol = min_balancedness_tol
self.honest = honest
self.n_jobs = n_jobs
self.random_state = random_state
self.verbose = verbose
self.warm_start = warm_start
self.max_samples = max_samples
[docs] def apply(self, X):
"""
Apply trees in the forest to X, return leaf indices.
Parameters
----------
X : array_like of shape (n_samples, n_features)
The input samples. Internally, it will be converted to
``dtype=np.float64``.
Returns
-------
X_leaves : ndarray of shape (n_samples, n_estimators)
For each datapoint x in X and for each tree in the forest,
return the index of the leaf x ends up in.
"""
X = self._validate_X_predict(X)
results = Parallel(n_jobs=self.n_jobs, verbose=self.verbose, backend="threading")(
delayed(tree.apply)(X, check_input=False)
for tree in self.estimators_)
return np.array(results).T
[docs] def decision_path(self, X):
"""
Return the decision path in the forest.
Parameters
----------
X : array_like of shape (n_samples, n_features)
The input samples. Internally, it will be converted to
``dtype=np.float64``.
Returns
-------
indicator : sparse matrix of shape (n_samples, n_nodes)
Return a node indicator matrix where non zero elements indicates
that the samples goes through the nodes. The matrix is of CSR
format.
n_nodes_ptr : ndarray of shape (n_estimators + 1,)
The columns from indicator[n_nodes_ptr[i]:n_nodes_ptr[i+1]]
gives the indicator value for the i-th estimator.
"""
X = self._validate_X_predict(X)
indicators = Parallel(n_jobs=self.n_jobs, verbose=self.verbose, backend='threading')(
delayed(tree.decision_path)(X, check_input=False)
for tree in self.estimators_)
n_nodes = [0]
n_nodes.extend([i.shape[1] for i in indicators])
n_nodes_ptr = np.array(n_nodes).cumsum()
return sparse_hstack(indicators).tocsr(), n_nodes_ptr
[docs] def fit(self, X, y, *, sample_weight=None, **kwargs):
"""
Build a forest of trees from the training set (X, y) and any other auxiliary variables.
Parameters
----------
X : array_like of shape (n_samples, n_features)
The training input samples. Internally, its dtype will be converted
to ``dtype=np.float64``.
y : array_like of shape (n_samples,) or (n_samples, n_treatments)
The outcome values for each sample and for each treatment.
sample_weight : array_like of shape (n_samples,), default None
Sample weights. If None, then samples are equally weighted. Splits
that would create child nodes with net zero or negative weight are
ignored while searching for a split in each node.
**kwargs : dictionary of array_like items of shape (n_samples, d_var)
Auxiliary random variables
Returns
-------
self : object
"""
X, y = check_X_y(X, y, multi_output=True)
if sample_weight is not None:
sample_weight = _check_sample_weight(sample_weight, X, DOUBLE)
# Remap output
n_samples, self.n_features_ = X.shape
y = np.atleast_1d(y)
if y.ndim == 1:
# reshape is necessary to preserve the data contiguity against vs
# [:, np.newaxis] that does not.
y = np.reshape(y, (-1, 1))
self.n_outputs_ = y.shape[1]
if getattr(y, "dtype", None) != DOUBLE or not y.flags.contiguous:
y = np.ascontiguousarray(y, dtype=DOUBLE)
if getattr(X, "dtype", None) != DTYPE:
X = X.astype(DTYPE)
# Get subsample sample size
n_samples_subsample = _get_n_samples_subsample(
n_samples=n_samples,
max_samples=self.max_samples
)
# Check parameters
self._validate_estimator()
random_state = check_random_state(self.random_state)
# We re-initialize the subsample_random_seed_ only if we are not in warm_start mode or
# if this is the first `fit` call of the warm start mode.
if (not self.warm_start) or (not hasattr(self, 'subsample_random_seed_')):
self.subsample_random_seed_ = random_state.randint(MAX_INT)
else:
random_state.randint(MAX_INT) # just advance random_state
subsample_random_state = check_random_state(self.subsample_random_seed_)
if not self.warm_start or not hasattr(self, "estimators_"):
# Free allocated memory, if any
self.estimators_ = []
self.slices_ = []
# the below are needed to replicate randomness of subsampling when warm_start=True
self.n_samples_ = []
self.n_samples_subsample_ = []
n_more_estimators = self.n_estimators - len(self.estimators_)
if n_more_estimators < 0:
raise ValueError('n_estimators=%d must be larger or equal to '
'len(estimators_)=%d when warm_start==True'
% (self.n_estimators, len(self.estimators_)))
elif n_more_estimators == 0:
warn("Warm-start fitting without increasing n_estimators does not "
"fit new trees.", UserWarning)
else:
if self.warm_start and len(self.estimators_) > 0:
# We draw from the random state to get the random state we
# would have got if we hadn't used a warm_start.
random_state.randint(MAX_INT, size=len(self.estimators_))
trees = [self._make_estimator(append=False,
random_state=random_state).init()
for i in range(n_more_estimators)]
if self.warm_start:
# Advancing subsample_random_state. Assumes each prior fit call has the same number of
# samples at fit time. If not then this would not exactly replicate a single batch execution,
# but would still advance randomness enough so that tree subsamples will be different.
for _, n_, ns_ in zip(range(len(self.estimators_)), self.n_samples_, self.n_samples_subsample_):
subsample_random_state.choice(n_, ns_, replace=False)
s_inds = [subsample_random_state.choice(n_samples, n_samples_subsample, replace=False)
for _ in range(n_more_estimators)]
# Parallel loop: we prefer the threading backend as the Cython code
# for fitting the trees is internally releasing the Python GIL
# making threading more efficient than multiprocessing in
# that case. However, for joblib 0.12+ we respect any
# parallel_backend contexts set at a higher level,
# since correctness does not rely on using threads.
trees = Parallel(n_jobs=self.n_jobs, verbose=self.verbose, backend='threading')(
delayed(t.fit)(X[s], y[s],
sample_weight=sample_weight[s] if sample_weight is not None else None,
check_input=False)
for t, s in zip(trees, s_inds))
# Collect newly grown trees
self.estimators_.extend(trees)
self.n_samples_.extend([n_samples] * len(trees))
self.n_samples_subsample_.extend([n_samples_subsample] * len(trees))
return self
[docs] def get_subsample_inds(self,):
""" Re-generate the example same sample indices as those at fit time using same pseudo-randomness.
"""
check_is_fitted(self)
subsample_random_state = check_random_state(self.subsample_random_seed_)
return [subsample_random_state.choice(n_, ns_, replace=False)
for n_, ns_ in zip(self.n_samples_, self.n_samples_subsample_)]
[docs] def feature_importances(self, max_depth=4, depth_decay_exponent=2.0):
"""
The feature importances based on the amount of parameter heterogeneity they create.
The higher, the more important the feature.
Parameters
----------
max_depth : int, default 4
Splits of depth larger than `max_depth` are not used in this calculation
depth_decay_exponent: double, default 2.0
The contribution of each split to the total score is re-weighted by 1 / (1 + `depth`)**2.0.
Returns
-------
feature_importances_ : ndarray of shape (n_features,)
Normalized total parameter heterogeneity inducing importance of each feature
"""
check_is_fitted(self)
all_importances = Parallel(n_jobs=self.n_jobs, backend='threading')(
delayed(tree.feature_importances)(
max_depth=max_depth, depth_decay_exponent=depth_decay_exponent)
for tree in self.estimators_ if tree.tree_.node_count > 1)
if not all_importances:
return np.zeros(self.n_features_, dtype=np.float64)
all_importances = np.mean(all_importances,
axis=0, dtype=np.float64)
return all_importances / np.sum(all_importances)
@property
def feature_importances_(self):
return self.feature_importances()
def _validate_X_predict(self, X):
"""
Validate X whenever one tries to predict, apply, and other predict methods."""
check_is_fitted(self)
return self.estimators_[0]._validate_X_predict(X, check_input=True)
[docs] def predict_value(self, X):
""" Predict the expected value of each treatment for each sample
Parameters
----------
X : {array_like} of shape (n_samples, n_features)
The input samples. Internally, it will be converted to
``dtype=np.float64``.
Returns
-------
welfare : array_like of shape (n_samples, n_treatments)
The conditional average welfare for each treatment for the group of each sample defined by the tree
"""
check_is_fitted(self)
# Check data
X = self._validate_X_predict(X)
# Assign chunk of trees to jobs
n_jobs, _, _ = _partition_estimators(self.n_estimators, self.n_jobs)
# avoid storing the output of every estimator by summing them here
y_hat = np.zeros((X.shape[0], self.n_outputs_), dtype=np.float64)
# Parallel loop
lock = threading.Lock()
Parallel(n_jobs=n_jobs, verbose=self.verbose, backend='threading', require="sharedmem")(
delayed(_accumulate_prediction)(e.predict_value, X, [y_hat], lock)
for e in self.estimators_)
y_hat /= len(self.estimators_)
return y_hat
[docs] def predict_proba(self, X):
""" Predict the probability of recommending each treatment
Parameters
----------
X : {array_like} of shape (n_samples, n_features)
The input samples. Internally, it will be converted to
``dtype=np.float64``.
check_input : bool, default True
Allow to bypass several input checking.
Don't use this parameter unless you know what you do.
Returns
-------
treatment_proba : array_like of shape (n_samples, n_treatments)
The probability of each treatment recommendation
"""
check_is_fitted(self)
# Check data
X = self._validate_X_predict(X)
# Assign chunk of trees to jobs
n_jobs, _, _ = _partition_estimators(self.n_estimators, self.n_jobs)
# avoid storing the output of every estimator by summing them here
y_hat = np.zeros((X.shape[0], self.n_outputs_), dtype=np.float64)
# Parallel loop
lock = threading.Lock()
Parallel(n_jobs=n_jobs, verbose=self.verbose, require="sharedmem")(
delayed(_accumulate_prediction)(e.predict_proba, X, [y_hat], lock)
for e in self.estimators_)
y_hat /= len(self.estimators_)
return y_hat
[docs] def predict(self, X):
""" Predict the best treatment for each sample
Parameters
----------
X : {array_like} of shape (n_samples, n_features)
The input samples. Internally, it will be converted to
``dtype=np.float64``.
Returns
-------
treatment : array_like of shape (n_samples)
The recommded treatment, i.e. the treatment index most often predicted to have the highest reward
for each sample. Recommended treatments are aggregated from each tree in the ensemble and the treatment
that receives the most votes is returned. Use `predict_proba` to get the fraction of trees in the ensemble
that recommend each treatment for each sample.
"""
return np.argmax(self.predict_proba(X), axis=1)