[MRG + 1] Fix gradient boosting overflow and various other float comparison on == (#7970)

chenhe95 · amueller · commit 919b4a8fbdf0 · 2017-03-07T09:21:38.000-05:00
* reintroduced isclose() and flake8 fixes to fixes.py

* changed == 0.0 to isclose(...)

* example changes

* changed back to abs() &lt; epsilon

* flake8 convention on file

* reverted flake8 fixes

* reverted flake8 fixes (2)

* np.finfo(np.float32).tiny instead of hard coded epsilon 1e-150

* reverted to 1e-150

* whats new modified
diff --git a/doc/whats_new.rst b/doc/whats_new.rst
@@ -185,6 +185,10 @@ Bug fixes
      ``download_if_missing`` keyword.  This was fixed in :issue:`7944` by
      :user:`Ralf Gommers <rgommers>`.
 
+   - Fixed a bug in :class:`sklearn.ensemble.GradientBoostingClassifier`
+     and :class:`sklearn.ensemble.GradientBoostingRegressor`
+     where a float being compared to ``0.0`` using ``==`` caused a divide by zero
+     error. This was fixed in :issue:`7970` by :user:`He Chen <chenhe95>`.
 
    - Fix a bug regarding fitting :class:`sklearn.cluster.KMeans` with a
      sparse array X and initial centroids, where X's means were unnecessarily
diff --git a/sklearn/ensemble/gradient_boosting.py b/sklearn/ensemble/gradient_boosting.py
@@ -510,7 +510,8 @@ def _update_terminal_region(self, tree, terminal_regions, leaf, X, y,
         numerator = np.sum(sample_weight * residual)
         denominator = np.sum(sample_weight * (y - residual) * (1 - y + residual))
 
-        if denominator == 0.0:
+        # prevents overflow and division by zero
+        if abs(denominator) < 1e-150:
             tree.value[leaf, 0, 0] = 0.0
         else:
             tree.value[leaf, 0, 0] = numerator / denominator
@@ -576,7 +577,8 @@ def _update_terminal_region(self, tree, terminal_regions, leaf, X, y,
         denominator = np.sum(sample_weight * (y - residual) *
                              (1.0 - y + residual))
 
-        if denominator == 0.0:
+        # prevents overflow and division by zero
+        if abs(denominator) < 1e-150:
             tree.value[leaf, 0, 0] = 0.0
         else:
             tree.value[leaf, 0, 0] = numerator / denominator
@@ -633,7 +635,8 @@ def _update_terminal_region(self, tree, terminal_regions, leaf, X, y,
         numerator = np.sum(y_ * sample_weight * np.exp(-y_ * pred))
         denominator = np.sum(sample_weight * np.exp(-y_ * pred))
 
-        if denominator == 0.0:
+        # prevents overflow and division by zero
+        if abs(denominator) < 1e-150:
             tree.value[leaf, 0, 0] = 0.0
         else:
             tree.value[leaf, 0, 0] = numerator / denominator
