def plot_kelbow_visualizer(self,
scorer='distortion',
axis=None,
plt_show=True):
return kelbow_visualizer(KMeans(),
self.data,
k=self.n_clusters,
metric=scorer,
ax=axis,
show=plt_show)
def test_quick_method_params(self):
"""
Test the quick method correctly consumes the user-provided parameters
"""
X, y = make_blobs(centers=3)
custom_title = "My custom title"
model = KMeans(3, random_state=13)
oz = kelbow_visualizer(model,
X,
sample_weight=np.ones(X.shape[0]),
title=custom_title)
assert oz.title == custom_title
def test_quick_method(self):
"""
Test the quick method producing a valid visualization
"""
X, y = make_blobs(n_samples=1000,
n_features=12,
centers=8,
shuffle=False,
random_state=2)
model = MiniBatchKMeans(3, random_state=43)
oz = kelbow_visualizer(model, X, random_state=13, legend=False)
assert isinstance(oz, KElbowVisualizer)
self.assert_images_similar(oz)
def __print_elbow(self, k, visualizer=None):
self.k = k
self.visualizer = kelbow_visualizer(KMeans(), self, self.k)
self.visualizer.show()
col=3)
fig.update_layout(yaxis_title='Value')
plot(fig)
# -------------------------------
# 2. Normalization
customers_data_norm = normalizer(df=customers_data_cleaned,
norm_type='StandardScaler')
# ***************************************************
# ****** Modelling ******
# ***************************************************
# -------------------------------
# 1. Extract optimize number of clusters based on the Elbow Method
kelbow_visualizer(KMeans(random_state=4), customers_data_norm, k=(2, 10))
# -------------------------------
# 2. Apply clustering model
clustering_model = KMeans(n_clusters=5, random_state=42)
clustering_model.fit(customers_data_norm)
customers_data_norm['cluster'] = clustering_model.labels_
customers_data_norm[
'cluster_str'] = 'cluster' + customers_data_norm['cluster'].astype(str)
customers_data['cluster'] = clustering_model.labels_
customers_data['cluster_str'] = 'cluster' + customers_data['cluster'].astype(
str)
# ***************************************************