8000 [MRG+1] Fix optics metric issues (DOC and precomputed) by adrinjalali · Pull Request #12028 · scikit-learn/scikit-learn · GitHub
[go: up one dir, main page]
Skip to content

[MRG+1] Fix optics metric issues (DOC and precomputed) #12028

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 7 commits into from
Sep 17, 2018
Merged

[MRG+1] Fix optics metric issues (DOC and precomputed) #12028

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
7 commits
Select commit Hold shift + click to select a range
b32d70f
Support precomputed distance matrix
adrinjalali Sep 6, 2018
414be06
sync docstrings
adrinjalali Sep 6, 2018
7df42ff
remove sparse
adrinjalali Sep 6, 2018
23bc0ac
Merge branch 'master' into optics/metric
adrinjalali Sep 9, 2018
b1437c0
put back reading from precomputed matrix
adrinjalali Sep 10, 2018
5b510ad
simplify test
adrinjalali Sep 10, 2018
56cba77
revert to float
adrinjalali Sep 10, 2018
File filter

Filter by extension

8000 Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
65 changes: 53 additions & 12 deletions sklearn/cluster/optics_.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -52,11 +52,30 @@ def optics(X, min_samples=5, max_eps=np.inf, metric='euclidean',
shorter run times.

metric : string or callable, optional (default='euclidean')
The distance metric to use for neighborhood lookups. Default is
"euclidean". Other options include "minkowski", "manhattan",
"chebyshev", "haversine", "seuclidean", "hamming", "canberra",
and "braycurtis". The "wminkowski" and "mahalanobis" metrics are
also valid with an additional argument.
metric to use for distance computation. Any metric from scikit-learn
or scipy.spatial.distance can be used.

If metric is a callable function, it is called on each
pair of instances (rows) and the resulting value recorded. The callable
should take two arrays as input and return one value indicating the
distance between them. This works for Scipy's metrics, but is less
efficient than passing the metric name as a string.

Distance matrices are not supported.
Valid values for metric are:

- from scikit-learn: ['cityblock', 'cosine', 'euclidean', 'l1', 'l2',
'manhattan']

- from scipy.spatial.distance: ['braycurtis', 'canberra', 'chebyshev',
'correlation', 'dice', 'hamming', 'jaccard', 'kulsinski',
'mahalanobis', 'minkowski', 'rogerstanimoto', 'russellrao',
'seuclidean', 'sokalmichener', 'sokalsneath', 'sqeuclidean',
'yule']

See the documentation for scipy.spatial.distance for details on these
metrics.

p : integer, optional (default=2)
Parameter for the Minkowski metric from
Expand Down Expand Up @@ -182,11 +201,30 @@ class OPTICS(BaseEstimator, ClusterMixin):
shorter run times.

metric : string or callable, optional (default='euclidean')
The distance metric to use for neighborhood lookups. Default is
"euclidean". Other options include "minkowski", "manhattan",
"chebyshev", "haversine", "seuclidean", "hamming", "canberra",
and "braycurtis". The "wminkowski" and "mahalanobis" metrics are
also valid with an additional argument.
metric to use for distance computation. Any metric from scikit-learn
or scipy.spatial.distance can be used.

If metric is a callable function, it is called on each
pair of instances (rows) and the resulting value recorded. The callable
should take two arrays as input and return one value indicating the
distance between them. This works for Scipy's metrics, but is less
efficient than passing the metric name as a string.

Distance matrices are not supported.

Valid values for metric are:

- from scikit-learn: ['cityblock', 'cosine', 'euclidean', 'l1', 'l2',
'manhattan']

- from scipy.spatial.distance: ['braycurtis', 'canberra', 'chebyshev',
'correlation', 'dice', 'hamming', 'jaccard', 'kulsinski',
'mahalanobis', 'minkowski', 'rogerstanimoto', 'russellrao',
'seuclidean', 'sokalmichener', 'sokalsneath', 'sqeuclidean',
'yule']

See the documentation for scipy.spatial.distance for details on these
metrics.

p : integer, optional (default=2)
Parameter for the Minkowski metric from
Expand Down Expand Up @@ -419,8 +457,11 @@ def _set_reach_dist(self, point_index, processed, X, nbrs):
# Everything is already processed. Return to main loop
return point_index

dists = pairwise_distances(P, np.take(X, unproc, axis=0),
self.metric, n_jobs=1).ravel()
if self.metric == 'precomputed':
dists = X[point_index, unproc]
else:
dists = pairwise_distances(P, np.take(X, unproc, axis=0),
self.metric, n_jobs=None).ravel()

rdists = np.maximum(dists, self.core_distances_[point_index])
new_reach = np.minimum(np.take(self.reachability_, unproc), rdists)
Expand Down
13 changes: 13 additions & 0 deletions sklearn/cluster/tests/test_optics.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -11,6 +11,7 @@
from sklearn.cluster.optics_ import _TreeNode, _cluster_tree
from sklearn.cluster.optics_ import _find_local_maxima
from sklearn.metrics.cluster import contingency_matrix
from sklearn.metrics.pairwise import pairwise_distances
from sklearn.cluster.dbscan_ import DBSCAN
from sklearn.utils.testing import assert_equal, assert_warns
from sklearn.utils.testing import assert_array_equal
Expand Down Expand Up @@ -436,3 +437,15 @@ def test_reach_dists():
else:
# we compare to truncated decimals, so use atol
assert_allclose(clust.reachability_, np.array(v), atol=1e-5)


def test_precomputed_dists():
redX = X[::10]
dists = pairwise_distances(redX, metric='euclidean')
clust1 = OPTICS(min_samples=10, algorithm='brute',
metric='precomputed').fit(dists)
clust2 = OPTICS(min_samples=10, algorithm='brute',
metric='euclidean').fit(redX)

assert_allclose(clust1.reachability_, clust2.reachability_)
assert_array_equal(clust1.labels_, clust2.labels_)
0