def kmeans(self, x, df):
for nc in att.NUMBER_CLUSTERS:
for mi in att.MAX_ITERATIONS:
for nr in att.NUMBER_INIT:
configuration, alg = clustering.kmeans(nc, nr, mi)
if results.path_exists(configuration) is False:
self.execute_clustering(alg, configuration, df, x)
def affinity_propagation(self, x, df):
for mi in att.MAX_ITERATIONS:
for nr in att.NUMBER_INIT:
configuration, alg = clustering.affinity_propagation(
nr, mi, att.DAMPING)
if results.path_exists(configuration) is False:
self.execute_clustering(alg, configuration, df, x)
def dbscan(self, x, df):
for metric in att.METRIC:
for eps in att.EPS:
for min_samples in att.MIN_SAMPLES:
configuration, alg = clustering.dbscan(
metric, eps, min_samples)
if results.path_exists(configuration) is False:
self.execute_clustering(alg, configuration, df, x)
def spectral(self, x, df):
for affinity in att.AFFINITY_SPECTRAL:
for labels in att.ASSIGN_LABELS:
for n_neighbors in att.NUMBER_NEIGHBORS:
for number_init in att.NUMBER_INIT:
configuration, alg = clustering.spectral_clustering(
affinity, labels, number_init, n_neighbors)
if results.path_exists(configuration) is False:
self.execute_clustering(alg, configuration, df, x)
def agglomerative(self, x, df):
for nc in att.NUMBER_CLUSTERS:
for affinity in att.METRIC:
for linkage in att.LINKAGE:
if linkage != "ward" and affinity != "euclidean":
configuration, alg = clustering.agglomerative(
nc, affinity, linkage)
if results.path_exists(configuration) is False:
self.execute_clustering(alg, configuration, df, x)