ENH Optimize runtime for IsolationForest (scikit-learn#23252)

Co-authored-by: Thomas J. Fan <[email protected]>

doc/whats_new/v1.2.rst

@@ -44,6 +44,12 @@ Changelog
 - |Enhancement| :class:`cluster.Birch` now preserves dtype for `numpy.float32`
  inputs. :pr:`22968` by `Meekail Zain <micky774>`.
 
+:mod:`sklearn.ensemble`
+.......................
+
+- |Efficiency| Improve runtime performance of :class:`ensemble.IsolationForest`
+ by avoiding data copies. :pr:`23252` by :user:`Zhehao Liu <MaxwellLZH>`.
+
 Code and Documentation Contributors
 -----------------------------------
 

sklearn/ensemble/_bagging.py

@@ -88,6 +88,8 @@ def _parallel_build_estimators(
  bootstrap_features = ensemble.bootstrap_features
  support_sample_weight = has_fit_parameter(ensemble.base_estimator_, "sample_weight")
  has_check_input = has_fit_parameter(ensemble.base_estimator_, "check_input")
+ requires_feature_indexing = bootstrap_features or max_features != n_features
+
  if not support_sample_weight and sample_weight is not None:
  raise ValueError("The base estimator doesn't support sample weight")
 
@@ -135,10 +137,11 @@ def _parallel_build_estimators(
  not_indices_mask = ~indices_to_mask(indices, n_samples)
  curr_sample_weight[not_indices_mask] = 0
 
- estimator_fit(X[:, features], y, sample_weight=curr_sample_weight)
-
+ X_ = X[:, features] if requires_feature_indexing else X
+ estimator_fit(X_, y, sample_weight=curr_sample_weight)
  else:
- estimator_fit(X[indices][:, features], y[indices])
+ X_ = X[indices][:, features] if requires_feature_indexing else X[indices]
+ estimator_fit(X_, y[indices])
 
  estimators.append(estimator)
  estimators_features.append(features)

sklearn/ensemble/_iforest.py

@@ -8,6 +8,7 @@
 from warnings import warn
 
 from ..tree import ExtraTreeRegressor
+from ..tree._tree import DTYPE as tree_dtype
 from ..utils import (
  check_random_state,
  check_array,
@@ -80,6 +81,9 @@ class IsolationForest(OutlierMixin, BaseBagging):
  - If int, then draw `max_features` features.
  - If float, then draw `max_features * X.shape[1]` features.
 
+ Note: using a float number less than 1.0 or integer less than number of
+ features will enable feature subsampling and leads to a longerr runtime.
+
  bootstrap : bool, default=False
  If True, individual trees are fit on random subsets of the training
  data sampled with replacement. If False, sampling without replacement
@@ -254,7 +258,7 @@ def fit(self, X, y=None, sample_weight=None):
  self : object
  Fitted estimator.
  """
- X = self._validate_data(X, accept_sparse=["csc"])
+ X = self._validate_data(X, accept_sparse=["csc"], dtype=tree_dtype)
  if issparse(X):
  # Pre-sort indices to avoid that each individual tree of the
  # ensemble sorts the indices.

sklearn/ensemble/tests/test_iforest.py

@@ -20,7 +20,7 @@
 from sklearn.ensemble import IsolationForest
 from sklearn.ensemble._iforest import _average_path_length
 from sklearn.model_selection import train_test_split
-from sklearn.datasets import load_diabetes, load_iris
+from sklearn.datasets import load_diabetes, load_iris, make_classification
 from sklearn.utils import check_random_state
 from sklearn.metrics import roc_auc_score
 
@@ -347,3 +347,13 @@ def test_n_features_deprecation():
 
  with pytest.warns(FutureWarning, match="`n_features_` was deprecated"):
  est.n_features_
+
+
+def test_iforest_with_n_jobs_does_not_segfault():
+ """Check that Isolation Forest does not segfault with n_jobs=2
+
+ Non-regression test for #23252
+ """
+ X, _ = make_classification(n_samples=85_000, n_features=100, random_state=0)
+ X = csc_matrix(X)
+ IsolationForest(n_estimators=10, max_samples=256, n_jobs=2).fit(X)