def test_get_intra_cluster_distances(self):
matrix = CondensedMatrix(CH_table1)
numpy.testing.assert_almost_equal(get_intra_cluster_distances(Cluster(None, [4,5]), matrix),[2.4494897427831779],5)
numpy.testing.assert_almost_equal(get_intra_cluster_distances(Cluster(None, [1,3,5]), matrix),[2.4494897427831779, 3.8729833462074170, 3.8729833462074170],5)
data = [0.0, 1.0, 1.0, 1.0,
1.0, 0.0, 0.0,
0.0, 0.0,
0.0]
matrix = CondensedMatrix(data)
expected_distance = 4
self.assertEqual(expected_distance, numpy.sum(get_intra_cluster_distances(Cluster(None, range(5)), matrix)))
def calculate_per_cluster_stats(best_clustering, matrix, parameters,
results_folder):
"""
CSV file
"""
file_name = parameters.get_value(
"file", default_value="per_cluster_stats") + ".csv"
stats_file_path = os.path.join(results_folder, file_name)
stats_file = open(stats_file_path, "w")
header_line = ","
for i in range(len(best_clustering.clusters)):
cluster = best_clustering.clusters[i]
header_line += "%s," % cluster.id
header_line = header_line[:-1] + "\n"
stats_file.write(header_line)
for i in range(len(best_clustering.clusters)):
cluster_i = best_clustering.clusters[i]
intra_distances = get_intra_cluster_distances(cluster_i, matrix)
radius = max(intra_distances) if intra_distances != [] else 0.
line = "%s(%.2f)," % (cluster_i.id, radius)
for j in range(0, i + 1):
line += ","
for j in range(i + 1, len(best_clustering.clusters)):
cluster_j = best_clustering.clusters[j]
line += "%.2f," % matrix[cluster_i.prototype, cluster_j.prototype]
line = line[:-1] + "\n"
stats_file.write(line)
stats_file.close()
return stats_file_path
def min_intracluster_distances(cls, clustering, matrix):
"""
Calculates d_min, the minimum internal distance.
@param clustering: The clustering being checked.
@param matrix: The condensed matrix containing all distances.
@return: d_min's value
"""
return numpy.min([
numpy.min(get_intra_cluster_distances(c, matrix))
for c in clustering.clusters
])
def ch_cluster_term(cls, cluster, global_mean_distance, matrix):
"""
Calculates one of the formula terms (ng-1)(D-d_g)
@param cluster: The cluster to use in calculation.
@param global_mean_distance: 'D'. Is the mean of the n*(n-1)/2 distances of all the elements.
@param matrix: The condensed matrix containing all distances.
@return: Calculated term.
"""
# Calculate cluster mean distance
n = len(cluster.all_elements)
cluster_mean_distance = mean(
numpy.array(get_intra_cluster_distances(cluster, matrix))**2)
return (n - 1) * (global_mean_distance - cluster_mean_distance)
def WGSS(cls, clusters, matrix):
"""
C-H description of the "Within group sum of squares".
@param clusters: An array with all clusters description (usually Clustering.clusters)
@param matrix: The condensed matrix containing all distances.
@return: The value of WGSS.
"""
wgss = 0
for c in clusters:
n = len(c.all_elements)
d = mean(numpy.array(get_intra_cluster_distances(c, matrix))**2)
wgss += (n - 1) * d
return wgss * 0.5
def WGSS(cls, clusters, matrix):
"""
C-H description of the "Within group sum of squares".
:param clusters: An array with all clusters description (usually Clustering.clusters)
:param matrix: The condensed matrix containing all distances.
:return: The value of WGSS.
"""
wgss = 0
for c in clusters:
n = len(c.all_elements)
d = mean(numpy.array(get_intra_cluster_distances( c, matrix))**2)
wgss += (n-1)*d
return wgss*0.5
def calculate_per_cluster_stats(best_clustering, matrix, parameters,
results_folder):
"""
CSV file
"""
file_name = parameters.get_value(
"file", default_value="per_cluster_stats") + ".csv"
stats_file_path = os.path.join(results_folder, file_name)
stats_file = open(stats_file_path, "w")
header_line = ","
for i in range(len(best_clustering.clusters)):
cluster = best_clustering.clusters[i]
header_line += "%s," % cluster.id
header_line = header_line[:-1] + "\n"
stats_file.write(header_line)
# TODO: Once clusterings and clusters become inmutable its medoids will be always updated,
# then this kind of operations will be unnecessary
update_medoids(best_clustering, matrix)
#----------------------------------------
for i in range(len(best_clustering.clusters)):
cluster_i = best_clustering.clusters[i]
try:
intra_distances = get_intra_cluster_distances(cluster_i, matrix)
diameter = max(intra_distances)
distances_from_proto = get_distances_of_elements_to(
cluster_i.prototype, cluster_i.all_elements, matrix)
radius = max(distances_from_proto)
except SingularClusterException:
diameter = 0
radius = 0
finally:
line = "%s(d: %.2f r: %.2f)," % (cluster_i.id, diameter, radius)
for j in range(0, i + 1):
line += ","
for j in range(i + 1, len(best_clustering.clusters)):
cluster_j = best_clustering.clusters[j]
line += "%.2f," % matrix[cluster_i.prototype, cluster_j.prototype]
line = line[:-1] + "\n"
stats_file.write(line)
stats_file.close()
return stats_file_path
def ch_cluster_term(cls, cluster, global_mean_distance, matrix):
"""
Calculates one of the formula terms (ng-1)(D-d_g)
:param cluster: The cluster to use in calculation.
:param global_mean_distance: 'D'. Is the mean of the n*(n-1)/2 distances of all the elements.
:param matrix: The condensed matrix containing all distances.
:return: Calculated term.
"""
# Calculate cluster mean distance
n = len(cluster.all_elements)
try:
cluster_mean_distance = mean(numpy.array(get_intra_cluster_distances( cluster, matrix))**2)
except SingularClusterException:
cluster_mean_distance = 0
return (n-1) * (global_mean_distance - cluster_mean_distance)
def calculate_per_cluster_stats(best_clustering, matrix, parameters, results_folder):
"""
CSV file
"""
file_name = parameters.get_value("file", default_value = "per_cluster_stats") + ".csv"
stats_file_path = os.path.join(results_folder,file_name)
stats_file = open(stats_file_path,"w")
header_line =","
for i in range(len(best_clustering.clusters)):
cluster = best_clustering.clusters[i]
header_line+="%s,"%cluster.id
header_line = header_line[:-1] +"\n"
stats_file.write(header_line)
# TODO: Once clusterings and clusters become inmutable its medoids will be always updated,
# then this kind of operations will be unnecessary
update_medoids(best_clustering, matrix)
#----------------------------------------
for i in range(len(best_clustering.clusters)):
cluster_i = best_clustering.clusters[i]
try:
intra_distances = get_intra_cluster_distances(cluster_i, matrix)
diameter = max(intra_distances)
distances_from_proto = get_distances_of_elements_to(cluster_i.prototype,
cluster_i.all_elements,
matrix)
radius = max(distances_from_proto)
except SingularClusterException:
diameter = 0
radius = 0
finally:
line = "%s(d: %.2f r: %.2f),"%(cluster_i.id, diameter, radius)
for j in range(0, i+1):
line += ","
for j in range(i+1, len(best_clustering.clusters)):
cluster_j = best_clustering.clusters[j]
line+="%.2f,"%matrix[ cluster_i.prototype, cluster_j.prototype]
line = line[:-1] + "\n"
stats_file.write(line)
stats_file.close()
return stats_file_path
def min_intracluster_distances(cls, clustering, matrix):
"""
Calculates d_min, the minimum internal distance.
@param clustering: The clustering being checked.
@param matrix: The condensed matrix containing all distances.
@return: d_min's value
"""
distances = []
for c in clustering.clusters:
try:
distances.append(numpy.min(get_intra_cluster_distances(c, matrix)))
except SingularClusterException:
# If we work with a singular cluster, we add 0s so that no min function
# fails. The convention for the distance of a cluster with only one element
# will be 0 in this case.
distances.append(0)
return numpy.min(distances)
def min_intracluster_distances(cls, clustering, matrix):
"""
Calculates d_min, the minimum internal distance.
@param clustering: The clustering being checked.
@param matrix: The condensed matrix containing all distances.
@return: d_min's value
"""
distances = []
for c in clustering.clusters:
try:
distances.append(
numpy.min(get_intra_cluster_distances(c, matrix)))
except SingularClusterException:
# If we work with a singular cluster, we add 0s so that no min function
# fails. The convention for the distance of a cluster with only one element
# will be 0 in this case.
distances.append(0)
return numpy.min(distances)
