DBCV_versions_comparison.py

# Piece of Python code developed by Davide Chicco on 17th November 2024
# developed for Python 3.11 on Linux Xubuntu 22
# released with creative commons license
# contact: [email protected]

import numpy as np
from permetrics import ClusteringMetric
from dbcv import dbcv


print("\n : : : : permetrics example : : : : ")

# data matrix
data = np.array([[1, 2],
                 [3, 4],
                 [5, 6],
                 [7, 8],
                 [9, 10]])

# labels array
labels = np.array([0, 0, 1, 1, 1])

# Calculate DBCV score
result = dbcv(X=data, y=labels)
print("FelSiq/DBCV =", result)

cm = ClusteringMetric(X=data, y_pred=labels)
print("permetrics dbcvi =", cm.DBCVI())


print("\n : : : : Stack Overflow question example : : : : ")
# https://stackoverflow.com/a/79188273

dataSO = np.array([[1, 1],       # Cluster 0
                   [1.2, 1.1],
                   [0.8, 1.2],
                   [4, 4],       # Cluster 1
                   [4.2, 4.1],
                   [3.8, 4.2]])

labelsSO = np.array([0, 0, 0, 1, 1, 1])

resultSO = dbcv(X=dataSO, y=labelsSO)
print("FelSiq/DBCV =", resultSO)

cmSO = ClusteringMetric(X=dataSO, y_pred=labelsSO)
print("permetrics dbcvi =", cmSO.DBCVI())


print("\n : : : : other example : : : : ")

# Generate random data for each cluster
cluster1 = np.random.normal(loc=0, scale=0.3, size=(25, 2))
cluster2 = np.random.normal(loc=3, scale=0.3, size=(25, 2))
cluster3 = np.random.normal(loc=6, scale=0.3, size=(25, 2))

# Combine the clusters into one dataset
example_data_user = np.vstack((cluster1, cluster2, cluster3))

# Create cluster labels
cluster_labels_user = np.array([1]*25 + [2]*25 + [3]*25)

resultUser = dbcv(X=example_data_user, y=cluster_labels_user)
print("FelSiq/DBCV =", resultUser)

cmUser = ClusteringMetric(X=example_data_user, y_pred=cluster_labels_user)
print("permetrics dbcvi =", cmUser.DBCVI(), "\n")