22

following the example Demo of DBSCAN clustering algorithm of Scikit Learning i am trying to store in an array the x, y of each clustering class 

import numpy as np
from sklearn.cluster import DBSCAN
from sklearn import metrics
from sklearn.datasets.samples_generator import make_blobs
from sklearn.preprocessing import StandardScaler
from pylab import *

# Generate sample data
centers = [[1, 1], [-1, -1], [1, -1]]
X, labels_true = make_blobs(n_samples=750, centers=centers, cluster_std=0.4, random_state=0)
X = StandardScaler().fit_transform(X) 

xx, yy = zip(*X)
scatter(xx,yy)
show()

enter image description here 

db = DBSCAN(eps=0.3, min_samples=10).fit(X)
core_samples = db.core_sample_indices_
labels = db.labels_
n_clusters_ = len(set(labels)) - (1 if -1 in labels else 0)
print n_clusters_
3

enter image description here

I'm trying to understand the DBSCAN implementation by scikit-learn, but from this point I'm having trouble. The number of cluster is 3 (n_clusters_) and I wish to store the x, y of each cluster in an array

Gianni Spear
  • 7,033
  • 22
  • 82
  • 131

2 Answers2

41

The first cluster is X[labels == 0], etc.:

clusters = [X[labels == i] for i in xrange(n_clusters_)]

and the outliers are

outliers = X[labels == -1]
Fred Foo
  • 355,277
  • 75
  • 744
  • 836
3

What do you mean by "of each cluster"?

In DBSCAN, clusters are not represented as centroids as in k-means, so there is no obvious representation of the cluster except its members. You already have the x and y position of the cluster members, as they are the input data.

So I'm not sure what the question is.

Andreas Mueller
  • 27,470
  • 8
  • 62
  • 74
  • hey Andreas, sorry if i was not clear. If you see in figure 2 each points (x, y) has a different color due to the cluster (e.g., 3 clusters in the example and back dots are noise). I wish to have an array with 3 blocks. Each block store the x, y points values of that cluter. – Gianni Spear Aug 14 '13 at 19:18
  • 1
    You have x and y position of the cluster members, but how can you know in which cluster these points are ? – user3378649 Apr 09 '14 at 22:43
  • 1
    @user3378649, there are lables corresponding to its cluster. `dbscan.labels_` – GoingMyWay Dec 17 '17 at 16:46