http://next.rikunabi.com/tech/docs/ct_s03600.jsp?p=002315

CodeIQ_ML_q2.py

#!/usr/bin/env python

import numpy as np
import scipy as sp
from scipy.cluster.vq import kmeans2,whiten
from pylab import *

s_data = np.genfromtxt('CodeIQ_eaten.txt', dtype='f8,f8,S1', delimiter=' ')
data = np.genfromtxt('CodeIQ_data.txt', delimiter=' ')

data = sp.r_[[[x[0],x[1]] for x in s_data],data]

whitened = whiten(data)
centroid, p_labels = kmeans2(whitened, k=3)

result = [[[],[]],[[],[]],[[],[]]]
for i,p_label in enumerate(p_labels):
	#print p_label,data[i]
	result[p_label][0] += [data[i][0]]
	result[p_label][1] += [data[i][1]]
#print result[0]

plot(result[0][0],result[0][1],'y.')
plot(result[1][0],result[1][1],'g.')
plot(result[2][0],result[2][1],'b.')

poisonous = np.array([[x[0],x[1]] for x in s_data if x[2]=='x'])
safemush = np.array([[x[0],x[1]] for x in s_data if x[2]=='o'])
plot(poisonous[:,0],poisonous[:,1], 'kx')
plot(safemush[:,0],safemush[:,1], 'ro')
show()

safe_id = [p_labels[i] for i,x in enumerate(s_data) if x[2]=='o']
#print safe_id

for safe in set(safe_id):
	for i in range(len(safe_id),len(result[safe][0])):
		print result[safe][0][i],result[safe][1][i]