In [9]:

    

%matplotlib inline
# pairwise distances looks at the distances between each
# record and finds closest members of the matrix.


    

In [10]:

    

from sklearn.metrics import pairwise
from sklearn.datasets import make_blobs
import numpy as np


    

In [11]:

    

points, labels = make_blobs()


    

In [12]:

    

distances = pairwise.pairwise_distances(points)


    

In [13]:

    

np.diag(distances)[:5]


    

    
Out[13]:





array([ 0.,  0.,  0.,  0.,  0.])

In [14]:

    

ranks = np.argsort(distances[0])


    

In [15]:

    

ranks[:5]


    

    
Out[15]:





array([ 0, 10, 68, 62, 15])

In [26]:

    

print points.shape
points[ranks][:5]


    

    




(100, 2)






    
Out[26]:





array([[ 0.39130467, -5.19677887],
       [ 0.49318055, -5.52736458],
       [ 0.71912691, -4.70887537],
       [ 0.23485245, -4.60406434],
       [ 0.76108915, -4.69760094]])

In [17]:

    

import matplotlib.pyplot as plt


    

In [25]:

    

f,ax = plt.subplots(figsize=(7,5))
ax.scatter(points[ranks][0], points[ranks][1])
ax.legend(loc='best')


    

In [ ]: