Merge pull request scikit-learn#6400 from scikit-learn/radneigh-njobs

[MRG + 1] ENH use n_jobs in brute-force radius_neighbors

Author: MechCoder

sklearn/neighbors/base.py

@@ -583,11 +583,12 @@ class from an array representing our data set and ask who's
  # for efficiency, use squared euclidean distances
  if self.effective_metric_ == 'euclidean':
  dist = pairwise_distances(X, self._fit_X, 'euclidean',
- squared=True)
+ n_jobs=self.n_jobs, squared=True)
  radius *= radius
  else:
  dist = pairwise_distances(X, self._fit_X,
  self.effective_metric_,
+ n_jobs=self.n_jobs,
  **self.effective_metric_params_)
 
  neigh_ind_list = [np.where(d <= radius)[0] for d in dist]