def elbow(X,
range_clusters=range(2, 6),
alg='kmeans',
cat_features=[],
random_state=42):
inertias = []
ks = range_clusters
model = None
for k in ks:
if alg == 'kmeans':
model = KMeans(n_clusters=k, random_state=random_state)
elif alg == 'kmodes':
model = KMeans(n_clusters=k, random_state=random_state)
else:
model = KPrototypes(n_clusters=k,
cat_features=cat_features,
random_state=random_state)
model.fit(X.values)
# centroids_, clusters_, inertia_ = k_means(X_final.values, k=k)
inertias.append(model.inertia)
plt.plot(ks, inertias, '-o', color='black')
plt.xlabel('number of clusters, k')
plt.ylabel('inertia')
plt.title(alg)
plt.xticks(ks)
plt.show()
def get_clusterer(n_clusters,
cat_features=[],
alg='kmeans',
agglo_params=None,
random_state=10):
clusterer = None
metric = ''
# if len(cat_features) == 0:
if alg == 'agglo':
clusterer = AgglomerativeClustering(affinity=agglo_params[0],
compute_full_tree='auto',
linkage=agglo_params[1],
memory=None,
n_clusters=n_clusters,
pooling_func='deprecated')
metric = 'euclidean'
elif alg == 'kmeans':
clusterer = KMeans(n_clusters=n_clusters, random_state=random_state)
metric = 'euclidean'
elif alg == 'kmodes':
clusterer = KModes(n_clusters=n_clusters, random_state=random_state)
metric = 'manhattan'
elif alg == 'fuzzy':
clusterer = FuzzyCMeans(n_clusters=n_clusters,
random_state=random_state)
metric = 'euclidean'
# else:
elif alg == 'kproto':
clusterer = KPrototypes(n_clusters=n_clusters,
cat_features=cat_features,
random_state=random_state)
metric = 'manhattan'
return clusterer, metric
# ## K-Prototypes
#
# In[27]:
from cluster.kprototypes import KPrototypes
best_clusters = None
best_centroids = None
best_r = None
best_score = -9999
cat_features = list(range(len(features)))
for r in range(25):
kp = KPrototypes(n_clusters=3, cat_features=cat_features, random_state=r)
kp.fit(df_original[features].values)
clusters = kp.labels
score = adjusted_rand_score(df_original[target], clusters)
if score > best_score:
best_clusters = clusters
best_centroids = kp.centroids
best_score = score
best_r = r
r_kproto = best_r
kprototypes_clusters = best_clusters
best_score
# In[28]:
# #### No PCA
#
# In[19]:
from cluster.kprototypes import KPrototypes
from cluster.metrics import get_metrics, rename_labels
from sklearn.metrics import adjusted_rand_score
best_clusters = None
best_centroids = None
best_r = None
best_score = -9999
cat_features = list(range(len(features)))
for r in range(25):
kp = KPrototypes(n_clusters=3, cat_features=cat_features, random_state=r)
kp.fit(df_original[features].values)
clusters = kp.labels
score = adjusted_rand_score(df_original[target], clusters)
if score > best_score:
best_clusters = clusters
best_centroids = kp.centroids
best_score = score
best_r = r
r_kproto = best_r
kprototypes_clusters = best_clusters
best_score
# In[20]:
# pca_df = pd.DataFrame(pca.transform(df_OHE[features_OHE]))
n_comp = 3
pca = PCA(n_components=n_comp)
X_scaled_pca = pca.fit_transform(X_scaled.values)
X_scaled_pca = pd.DataFrame(X_scaled_pca)
X_scaled_pca.head()
# In[138]:
graph_components(X_scaled, n_components=n_comp)
# In[139]:
from cluster.kprototypes import KPrototypes
cat_features = list(range(2, len(X_scaled.columns)))
kp = KPrototypes(n_clusters=3, cat_features=cat_features, random_state=5)
kp.fit(X_scaled.values)
plt.scatter(X_scaled_pca.values[:, 0],
X_scaled_pca.values[:, 1],
c=kp.labels,
s=50,
cmap='viridis')
centroids_pca = PCA(n_components=n_comp).fit_transform(kp.centroids)
plt.scatter(centroids_pca[:, 0], centroids_pca[:, 1], marker='x', c='r', s=200)
plt.title('K-prototypes')
plt.show()
# In[140]:
from decomposition.pca import PCA
print(f'Shape X: {X.values.shape}')
# print(f'Shape W: {W.shape}')
pca = PCA(n_components=3, random_state=5)
X_pca = pca.fit_transform(X.values)
print(f'Shape X_pca: {X_pca.shape}')
# In[19]:
from cluster.kprototypes import KPrototypes
# cat_features = list(range(2, len(X.columns)))
kp = KPrototypes(n_clusters=3, cat_features=[], random_state=8)
kp.fit(X_pca)
plt.scatter(X_pca[:, 0], X_pca[:, 1], c=kp.labels, s=50, cmap='viridis')
centroids_pca = PCA(n_components=3).fit_transform(kp.centroids)
plt.scatter(centroids_pca[:, 0], centroids_pca[:, 1], marker='x', c='r', s=200)
plt.title('K-prototypes')
plt.show()
# **Metrics**
# In[20]:
from cluster.metrics import get_metrics
range_n_clusters = [2, 3, 4]
silhouette(X_cat, X_cat_pca, alg='kmodes', range_clusters=range_n_clusters)
# **K-prototypes**
# In[110]:
# best random state
from cluster.kprototypes import KPrototypes
best_clusters = None
best_centroids = None
best_r = None
best_score = -9999
for r in range(20):
kp = KPrototypes(n_clusters=2, random_state=r)
kp.fit(X_num_scaled.values)
score = adjusted_rand_score(y, kp.labels)
if score > best_score:
best_clusters = kme.labels
best_centroids = kme.centroids
best_score = score
best_r = r
fcm_clusters = best_clusters
print('Best score:', best_score)
print('Best random state value:', best_r)
# In[111]:
# Visaulization k-prototypes
from cluster.kprototypes import KPrototypes