test_cluster.py

这是一个关于c-cluster的技术文章,它详细地介绍了cluster的一些技术分类及算法等,是学习聚类的入门好教程.

                    [ 1.7, 1.9 ],
                    [ 5.7, 5.9 ],
                    [ 5.7, 5.9 ],
                    [ 3.1, 3.3 ],
                    [ 5.4, 5.3 ],
                    [ 5.1, 5.5 ],
                    [ 5.0, 5.5 ],
                    [ 5.1, 5.2 ]])
  mask2 = array([[ 1, 1 ],
                 [ 1, 1 ],
                 [ 1, 1 ],
                 [ 1, 1 ],
                 [ 1, 1 ],
                 [ 1, 1 ],
                 [ 1, 1 ],
                 [ 1, 1 ],
                 [ 1, 1 ],
                 [ 1, 1 ],
                 [ 1, 1 ],
                 [ 1, 1 ],
                 [ 1, 1 ]])
  # Cluster assignments
  c1 = [0]
  c2 = [1,2]
  c3 = [3]

  print "First data set:"
  print_matrix(data1, mask1)
  print "Clusters are", c1, c2, c3
  distance = clusterdistance(data1, mask=mask1, weight=weight1, index1=c1, index2=c2, dist='e', method='a', transpose=0);
  print "Distance between cluster", c1, "and", c2, "is %7.3f." % distance
  distance = clusterdistance(data1, mask=mask1, weight=weight1, index1=c1, index2=c3, dist='e', method='a', transpose=0);
  print "Distance between cluster", c1, "and", c3, "is %7.3f." % distance
  distance = clusterdistance(data1, mask=mask1, weight=weight1, index1=c2, index2=c3, dist='e', method='a', transpose=0);
  print "Distance between cluster", c2, "and", c3, "is %7.3f." % distance

  # Cluster assignments
  c1 = [ 0, 1, 2, 3 ]
  c2 = [ 4, 5, 6, 7 ]
  c3 = [ 8 ]
  print "Second data set:"
  print_matrix(data2, mask2)
  print "Clusters are", c1, c2, c3
  distance = clusterdistance(data2, mask=mask2, weight=weight2, index1=c1, index2=c2, dist='e', method='a', transpose=0);
  print "Distance between cluster", c1, "and", c2, "is %7.3f." % distance
  distance = clusterdistance(data2, mask=mask2, weight=weight2, index1=c1, index2=c3, dist='e', method='a', transpose=0);
  print "Distance between cluster", c1, "and", c3, "is %7.3f." % distance
  distance = clusterdistance(data2, mask=mask2, weight=weight2, index1=c2, index2=c3, dist='e', method='a', transpose=0);
  print "Distance between cluster", c2, "and", c3, "is %7.3f." % distance
  print

def test_treecluster(module):
  if module=='Bio.Cluster':
    from Bio.Cluster import treecluster
  elif module=='Pycluster':
    from Pycluster import treecluster
  else:
    raise 'Unknown module name', module
  print "test_treecluster:"
  # First data set
  weight1 =  [ 1,1,1,1,1 ]
  data1   =  array([[  1.1, 2.2, 3.3, 4.4, 5.5], 
                    [  3.1, 3.2, 1.3, 2.4, 1.5], 
                    [  4.1, 2.2, 0.3, 5.4, 0.5], 
                    [ 12.1, 2.0, 0.0, 5.0, 0.0]])
  mask1 = array([[ 1, 1, 1, 1, 1], 
                 [ 1, 1, 1, 1, 1], 
                 [ 1, 1, 1, 1, 1], 
                 [ 1, 1, 1, 1, 1]])

  # Second data set
  weight2 =  [ 1,1 ]
  data2 = array([[ 0.8223, 0.9295 ],
                 [ 1.4365, 1.3223 ],
                 [ 1.1623, 1.5364 ],
                 [ 2.1826, 1.1934 ],
                 [ 1.7763, 1.9352 ],
                 [ 1.7215, 1.9912 ],
                 [ 2.1812, 5.9935 ],
                 [ 5.3290, 5.9452 ],
                 [ 3.1491, 3.3454 ],
                 [ 5.1923, 5.3156 ],
                 [ 4.7735, 5.4012 ],
                 [ 5.1297, 5.5645 ],
                 [ 5.3934, 5.1823 ]])
  mask2 = array([[ 1, 1 ],
                 [ 1, 1 ],
                 [ 1, 1 ],
                 [ 1, 1 ],
                 [ 1, 1 ],
                 [ 1, 1 ],
                 [ 1, 1 ],
                 [ 1, 1 ],
                 [ 1, 1 ],
                 [ 1, 1 ],
                 [ 1, 1 ],
                 [ 1, 1 ],
                 [ 1, 1 ]])

  # test first data set
  print "First data set:"
  print_matrix(data1, mask1)
  print "Pairwise average-linkage clustering"
  result, linkdist = treecluster(data=data1, mask=mask1, weight=weight1, transpose=0, method='a', dist='e')
  print "Number of nodes is %d (should be %d)" % (len(result), len(data1)-1)
  print "Number of link distances is %d (should be %d)" % (len(linkdist), len(data1)-1)
  for i in range(len(result)):
    print "Node %3d joins node %3d with node %3d; link distance is %7.3f" % (i, result[i][0], result[i][1], linkdist[i])

  print "Pairwise single-linkage clustering"
  result, linkdist = treecluster(data=data1, mask=mask1, weight=weight1, transpose=0, method='s', dist='e')
  print "Number of nodes is %d (should be %d)" % (len(result), len(data1)-1)
  print "Number of link distances is %d (should be %d)" % (len(linkdist), len(data1)-1)
  for i in range(len(result)):
    print "Node %3d joins node %3d with node %3d; link distance is %7.3f" % (i, result[i][0], result[i][1], linkdist[i])

  print "Pairwise centroid-linkage clustering"
  result, linkdist = treecluster(data=data1, mask=mask1, weight=weight1, transpose=0, method='c', dist='e')
  print "Number of nodes is %d (should be %d)" % (len(result), len(data1)-1)
  print "Number of link distances is %d (should be %d)" % (len(linkdist), len(data1)-1)
  for i in range(len(result)):
    print "Node %3d joins node %3d with node %3d; link distance is %7.3f" % (i, result[i][0], result[i][1], linkdist[i])

  print "Pairwise maximum-linkage clustering"
  result, linkdist = treecluster(data=data1, mask=mask1, weight=weight1, transpose=0, method='m', dist='e')
  print "Number of nodes is %d (should be %d)" % (len(result), len(data1)-1)
  print "Number of link distances is %d (should be %d)" % (len(linkdist), len(data1)-1)
  for i in range(len(result)):
    print "Node %3d joins node %3d with node %3d; link distance is %7.3f" % (i, result[i][0], result[i][1], linkdist[i])

  # Test second data set
  print "Second data set:"
  print "Pairwise average-linkage clustering"
  result, linkdist = treecluster(data=data2, mask=mask2, weight=weight2, transpose=0, method='a', dist='e')
  print "Number of nodes is %d (should be %d)" % (len(result), len(data2)-1)
  print "Number of link distances is %d (should be %d)" % (len(linkdist), len(data2)-1)
  for i in range(len(result)):
    print "Node %3d joins node %3d with node %3d; link distance is %7.3f" % (i, result[i][0], result[i][1], linkdist[i])

  print "Pairwise single-linkage clustering"
  result, linkdist = treecluster(data=data2, mask=mask2, weight=weight2, transpose=0, method='s', dist='e')
  print "Number of nodes is %d (should be %d)" % (len(result), len(data2)-1)
  print "Number of link distances is %d (should be %d)" % (len(linkdist), len(data2)-1)
  for i in range(len(result)):
    print "Node %3d joins node %3d with node %3d; link distance is %7.3f" % (i, result[i][0], result[i][1], linkdist[i])

  print "Pairwise centroid-linkage clustering"
  result, linkdist = treecluster(data=data2, mask=mask2, weight=weight2, transpose=0, method='c', dist='e')
  print "Number of nodes is %d (should be %d)" % (len(result), len(data2)-1)
  print "Number of link distances is %d (should be %d)" % (len(linkdist), len(data2)-1)
  for i in range(len(result)):
    print "Node %3d joins node %3d with node %3d; link distance is %7.3f" % (i, result[i][0], result[i][1], linkdist[i])

  print "Pairwise maximum-linkage clustering"
  result, linkdist = treecluster(data=data2, mask=mask2, weight=weight2, transpose=0, method='m', dist='e')
  print "Number of nodes is %d (should be %d)" % (len(result), len(data2)-1)
  print "Number of link distances is %d (should be %d)" % (len(linkdist), len(data2)-1)
  for i in range(len(result)):
    print "Node %3d joins node %3d with node %3d; link distance is %7.3f" % (i, result[i][0], result[i][1], linkdist[i])
  print

def test_somcluster(module):
  if module=='Bio.Cluster':
    from Bio.Cluster import somcluster
  elif module=='Pycluster':
    from Pycluster import somcluster
  else:
    raise 'Unknown module name', module
  print "test_somcluster:"

  # First data set
  weight1 = [ 1,1,1,1,1 ]
  data1 = array([[  1.1, 2.2, 3.3, 4.4, 5.5], 
                 [  3.1, 3.2, 1.3, 2.4, 1.5], 
                 [  4.1, 2.2, 0.3, 5.4, 0.5], 
                 [ 12.1, 2.0, 0.0, 5.0, 0.0]])
  mask1 = array([[ 1, 1, 1, 1, 1], 
                 [ 1, 1, 1, 1, 1], 
                 [ 1, 1, 1, 1, 1], 
                 [ 1, 1, 1, 1, 1]])

  # Second data set
  weight2 =  [ 1,1 ]
  data2 = array([[ 1.1, 1.2 ],
                 [ 1.4, 1.3 ],
                 [ 1.1, 1.5 ],
                 [ 2.0, 1.5 ],
                 [ 1.7, 1.9 ],
                 [ 1.7, 1.9 ],
                 [ 5.7, 5.9 ],
                 [ 5.7, 5.9 ],
                 [ 3.1, 3.3 ],
                 [ 5.4, 5.3 ],
                 [ 5.1, 5.5 ],
                 [ 5.0, 5.5 ],
                 [ 5.1, 5.2 ]])
  mask2 = array([[ 1, 1 ],
                 [ 1, 1 ],
                 [ 1, 1 ],
                 [ 1, 1 ],
                 [ 1, 1 ],
                 [ 1, 1 ],
                 [ 1, 1 ],
                 [ 1, 1 ],
                 [ 1, 1 ],
                 [ 1, 1 ],
                 [ 1, 1 ],
                 [ 1, 1 ],
                 [ 1, 1 ]])

  print "First data set:"
  clusterid, celldata = somcluster(data=data1, mask=mask1, weight=weight1, transpose=0, nxgrid=10, nygrid=10, inittau=0.02, niter=100, dist='e')
  print "Number of cluster ids is %d (should be %d)" % (len(clusterid), len(data1))
  print "Grid is %d-dimensional (should be 2-dimensional)" % len(clusterid[0])

  print "Second data set:"
  clusterid, celldata = somcluster(data=data2, mask=mask2, weight=weight2, transpose=0, nxgrid=10, nygrid=10, inittau=0.02, niter=100, dist='e')
  print "Number of cluster ids is %d (should be %d)" % (len(clusterid), len(data2))
  print "Grid is %d-dimensional (should be 2-dimensional)" % len(clusterid[0])
  print

def run_tests(module="Pycluster"):
  if module==[]: module = "Bio.Cluster"
  test_mean_median(module)
  test_matrix_parse(module)
  test_kcluster(module)
  test_clusterdistance(module)
  test_treecluster(module)
  test_somcluster(module)