def __init__(self, clusters: np.ndarray, classes: np.ndarray, aic: float,
bic: float, likelihood: float):
"""
:param clusters: The cluster assignments
:param classes: The class assignments
:param aic: The Akaike information criterion
:param bic: The Bayesian information criterion
:param likelihood: The log likelihood
"""
super().__init__(clusters, aic, bic, likelihood)
self.N = len(classes)
self.C = len(np.unique(classes))
self.purity = purity(clusters, classes)
self.purity2 = purity2(clusters, classes)
self.rand_index = rand_index(clusters, classes)
self.class_entropy = entropy(classes)
self.precision = precision(clusters, classes)
self.recall = recall(clusters, classes)
self.f1_measure = f1_measure(clusters, classes)
self.homogeneity = homogeneity(clusters, classes)
self.completeness = completeness(clusters, classes)
self.v_measure = v_measure(clusters, classes)
self.nv_measure = nv_measure(clusters, classes)
self.mutual_information = mutual_information(clusters, classes)
self.normalized_mutual_information = normalized_mutual_information(
clusters, classes)
self.normalized_mutual_information2 = normalized_mutual_information2(
clusters, classes)
def __init__(self, clusters: np.ndarray, aic: float, bic: float,
likelihood: float):
"""
:param clusters: The cluster assignments
:param aic: The Akaike information criterion
:param bic: The Bayesian information criterion
:param likelihood: The log likelihood
"""
self.aic = aic
self.bic = bic
self.likelihood = likelihood
self.K = len(np.unique(clusters))
self.cluster_entropy = entropy(clusters)
def test_entropy_multiple_items(self):
labels = [1, 2, 3, 1]
assert_almost_equal(cm.entropy(labels),
scipy.stats.entropy([0.5, 0.25, 0.25]))
def test_entropy_two_item(self):
labels = [1, 2]
assert_almost_equal(cm.entropy(labels), scipy.stats.entropy([0.5,
0.5]))
def test_entropy_one_item(self):
labels = [50]
assert_almost_equal(cm.entropy(labels), scipy.stats.entropy([1]))
def test_entropy_empty_set(self):
labels = []
assert_almost_equal(cm.entropy(labels), scipy.stats.entropy([]))